Same Blood Sugar, Different A1c – What a Cutoff Does and Does Not Tell You
Danielle Izaak-Lewis is a nutrition consultant, integrative and functional nutrition expert, and trauma-informed clinical nutrition strategist. She founded Mindful Nutrition & Wellness Solutions, helping high-achieving Black, Indigenous, and women of color recalibrate their health through evidence-based nutrition and cultural food wisdom.
Your fasting numbers have held steady for months. If you wear a glucose monitor, the daily average looks ordinary. You track, you log, you eat the way the guidelines say to eat. Then your hemoglobin A1c result posts a few tenths of a percentage point above the line, and the word on your chart changes. Nothing you measured yourself changed with it.
The gap between what you track and what the cutoff reports is not a verdict on your discipline. It reflects where someone drew the line, who stood in the reference group, and what the test measures.
My previous piece for Brainz, The Labs Look Normal, But Your Body is Still Under Siege, asked what a clean panel was never calibrated to detect. This piece asks about the numbers a panel does measure. Where does the line sit? Who was it drawn around? What does it cost to move it?
Lab Reference Ranges Weren’t Built on Every Body covers the thyroid case. A 2025 study recalculated thyroid reference ranges by age, sex, and race and reclassified 48.5 percent of people previously classified as having subclinical hypothyroidism as having normal thyroid function. This piece looks at two other tests, hemoglobin A1c and hemoglobin.
One glucose, two verdicts
Guidelines use a standard A1c cutoff when A1c is appropriate for diagnosis. A 2019 analysis of national survey data by Ford and colleagues compared A1c with glucose results from oral glucose tolerance testing in 5,324 Black and white adults. The A1c cutoff that best distinguished adults with diabetes from those without it was 6.0 percent in Black adults and 5.7 percent in white adults. For dysglycemia, the broader category, the figures were 5.7 percent and 5.5 percent, respectively. The authors concluded that A1c needs to be interpreted in relation to an individual’s glucose levels when used to diagnose diabetes or prediabetes.
A 2023 Kaiser Permanente Northern California study by Karter and colleagues looked at the relationship using continuous glucose monitors. The study included 1,788 patients with diabetes and used about 124 million readings to calculate mean glucose over the 90 days before each A1c test. For a given mean glucose, A1c ran 0.33 percentage points higher in African American patients than in white patients. Same glucose. Different A1c. Asian, Latino, and multiethnic patients did not differ significantly from white patients.
Then the authors reported the number that matters more. Sixty-five percent of the variation in the relationship between glucose and A1c was within racial and ethnic groups, and 9 percent was between them. The authors concluded that treatment decisions based on A1c targets need to be individualized and supported by direct glucose measurements.
One distinction matters here. Karter’s patients already had a diabetes diagnosis, so that study speaks to monitoring and targets. Ford’s study speaks to diagnosis.
The reason is still open
Three possible explanations emerge, and the studies do not agree. Red blood cell lifespan comes first. Cohen and colleagues (2008) labeled red cells in 12 people and found that cell lifespan varied enough among hematologically normal adults to shift A1c at the same average glucose. That study did not compare racial groups. It shows the mechanism exists. It does not show that the mechanism explains this gap. A 2021 review by Harris and colleagues adds that red cell survival, inherited hemoglobin variants such as those associated with sickling, and glucose-6-phosphate dehydrogenase deficiency can influence A1c. The review notes that reported effects in African American populations are contradictory and hard to explain.
The second explanation is how hemoglobin itself glycates. Bergenstal and colleagues (2017) followed 104 Black and 104 white people with type 1 diabetes for 12 weeks using continuous glucose monitoring. A1c was 0.4 percentage points higher in Black participants at a given mean glucose. Glycated albumin and fructosamine, two markers that do not depend on red cells, showed no significant racial differences in their relationships with mean glucose. The authors pointed to possible differences in hemoglobin glycation and noted that race explained only part of the difference in A1c.
The third explanation is glycemia that fasting tests miss. Selvin and colleagues (2011) studied 1,719 adults in the Atherosclerosis Risk in Communities Study and found that glycated albumin and fructosamine also ran higher in Black adults, even after adjusting for fasting glucose. Those markers do not depend on red cells, so the authors raised the possibility that glucose levels differed at times a fasting test would miss. The study did not use continuous glucose monitoring to test that possibility.
Each study holds part of the picture. None closes it.
A cutoff is a trade, and hemoglobin shows the price
Clinicians use hemoglobin to identify anemia. Sekhar and colleagues (2017) evaluated a historical cutoff for Black women that was 1.0 gram per deciliter (g/dL) lower than the standard cutoff. The World Health Organization’s 2024 guideline uses a cutoff of 12.0 g/dL for nonpregnant women ages 15 to 65 and does not recommend adjusting hemoglobin concentrations for race or ethnicity when assessing anemia in populations.
Sekhar and colleagues tested how hemoglobin cutoffs performed against an estimate of body iron calculated from ferritin and soluble transferrin receptor levels. The sample included 6,602 nonpregnant girls and women ages 12 to 49 from a national survey. Anemia can develop late in iron deficiency, so screening with hemoglobin alone can miss iron deficiency before anemia develops.
At the standard cutoff of 12.0 g/dL, hemoglobin caught 42.9 percent of iron deficiency cases in the full sample and wrongly flagged 4.5 percent of those without iron deficiency. Among Black participants, the historical cutoff of 11.0 g/dL caught 37.4 percent of cases and wrongly flagged 1.3 percent of those without iron deficiency.
The researchers then looked for the cutoffs that best balanced detection and false alarms. For the full sample, the cutoff was 12.8 g/dL. It caught 71.3 percent of cases and flagged 20.7 percent of participants without iron deficiency. For Black participants, the cutoff was 12.5 g/dL, with 80.7 percent of cases caught and 26.7 percent of those without iron deficiency flagged. For participants who were not Black, the cutoff was 13.0 g/dL, with 69.9 percent of cases caught and 20.7 percent of those without iron deficiency flagged.
Raising the line catches more cases and produces more false alarms. The authors suggested that a practical cutoff might lie between 12.0 and 12.8 g/dL and proposed analyzing cost-effectiveness as the next step. They left open whether the difference between 12.5 and 13.0 g/dL was clinically meaningful.
Two limits keep the picture honest. The study had no data on sickle cell disease or thalassemia, and the body iron estimate used as the reference was not adjusted for age or race. Age also shifted performance. Hemoglobin was better at distinguishing participants with iron deficiency from those without it at ages 22 to 49 (0.87) than at ages 12 to 21 (0.77). On that scale, 0.5 is a coin flip and 1.0 is perfect.
What context adds
Read together, these studies show one pattern three ways. A cutoff is a judgment about which error costs less, and that judgment changes with who stands in front of it. Race appears in these studies as a marker. Karter and colleagues describe race as a social construct, and they found much more variation in the relationship between glucose and A1c within racial and ethnic groups than between them. In Sekhar’s study, hemoglobin’s ability to detect iron deficiency varied by age, and the cutoffs that best balanced detection and false alarms differed across the groups studied. A cutoff works or fails depending on what else is known about the person holding the result.
That context is specific. A glucose pattern next to the A1c. Ferritin next to the hemoglobin. Your age, any inherited hemoglobin variant, and your own trend over time.
Bring three questions to your care team. What glucose measurement sits next to this A1c? What iron measure sits next to this hemoglobin? How does this result compare with my own earlier results as well as with the printed line? These questions inform a conversation with your clinician. They do not replace one.
The patterns you track are part of the picture. A cutoff is a judgment about a line, and that judgment improves when it meets the context you carry.
Read this alongside Lab Reference Ranges Weren’t Built on Every Body, which lays out the thyroid case with its figures and citations, and The Labs Look Normal, But Your Body is Still Under Siege, which asks what a clean panel was never built to detect.
Visit my website for more info!
Read more from Danielle Izaak-Lewis
Danielle Izaak-Lewis, Clinical Nutrition Strategist & Consultant
Danielle Izaak-Lewis is a nutrition consultant, integrative and functional nutrition expert, and trauma-informed clinical nutrition strategist. She founded Mindful Nutrition & Wellness Solutions to help high-achieving Black, Indigenous, and women of color recalibrate their health instead of following generic meal plans. Her work blends evidence-based nutrition science, epigenetics, and trauma-informed care with the cultural food wisdom many clients were taught to set aside. Danielle holds a Master's in Clinical Nutrition and built her practice after her own experience of stress debt revealed how those patterns show up in the body long before they show up as a diagnosis. She believes individual healing is generational healing.
References:
Ford CN et al. “Racial differences in performance of HbA1c for the classification of diabetes and prediabetes among US adults of non-Hispanic black and white race.” Diabetic Medicine (2019).
Karter AJ, Parker MM, Moffet HH, and Gilliam LK. “Racial and Ethnic Differences in the Association Between Mean Glucose and Hemoglobin A1c.” Diabetes Technology & Therapeutics (2023).
Cohen RM et al. “Red cell life span heterogeneity in hematologically normal people is sufficient to alter HbA1c.” Blood (2008).
Harris NS, Weaver KD, Beal SG, and Winter WE. “The Interaction between Hb A1C and Selected Genetic Factors in the African American Population in the USA.” The Journal of Applied Laboratory Medicine (2021).
Bergenstal RM et al. “Racial Differences in the Relationship of Glucose Concentrations and Hemoglobin A1c Levels.” Annals of Internal Medicine (2017).
Selvin E et al. “Racial differences in glycemic markers: a cross-sectional analysis of community-based data.” Annals of Internal Medicine (2011).
Sekhar DL, Kunselman AR, Chuang CH, and Paul IM. “Optimizing hemoglobin thresholds for detection of iron deficiency among reproductive-age women in the United States.” Translational Research (2017, published online in 2016).
Li Q et al. “Thyroid Function Reference Intervals by Age, Sex, and Race: A Cross-Sectional Study.” Annals of Internal Medicine (2025).
World Health Organization. Guideline on haemoglobin cutoffs to define anaemia in individuals and populations (2024). WHO guideline
