Keywords: HbA1c, Continuous Glucose Monitoring, Glycaemic Variability, Time in Range, Metabolic Health, Longevity, Diabetes Prevention, Precision Medicine
Introduction
Diabetes mellitus and its precursor states, prediabetes and insulin resistance represent a global pandemic of unprecedented scale. According to the International Diabetes Federation, more than 537 million adults live with diabetes worldwide, a figure projected to exceed 780 million by 2045. For decades, haemoglobin A1c (HbA1c) has occupied the central position in glycaemic monitoring, serving as the primary diagnostic criterion, the dominant treatment target, and the principal surrogate endpoint in landmark cardiovascular outcomes trials. Its clinical utility is well-established: HbA1c reflects the average glucose level over the preceding two to three months by measuring the proportion of glycated haemoglobin within circulating erythrocytes [1,2].
Yet the landscape of glycaemic science has shifted dramatically. A growing body of evidence has exposed fundamental limitations of HbA1c as a standalone metric, limitations rooted in the biology of erythrocyte turnover, the heterogeneity of haemoglobin variants across ethnic populations, and, most critically, the metric’s complete inability to capture glycaemic variability. Two patients may present with an identical HbA1c of 7.0% while experiencing profoundly different glycaemic patterns: one with stable, well-controlled glucose and another oscillating between dangerous hyperglycaemic excursions and hypoglycaemic episodes. HbA1c cannot distinguish between them [3,4].
Continuous glucose monitoring (CGM) technology has provided the clinical and scientific community with an unprecedented window into real-world glucose dynamics. Through metrics including Time in Range (TIR), Time Below Range (TBR), glucose variability measured by coefficient of variation (CV), and the Glucose Management Indicator (GMI), CGM offers a multidimensional portrait of glycaemic health that no single number can replicate. International expert consensus, including the 2019 International Consensus on Time in Range has endorsed CGM-derived metrics as clinically meaningful endpoints, and the American Diabetes Association (ADA) 2025 Standards of Care now formally recommend CGM integration across a broader population of patients with diabetes [2,5].
Beyond established diabetes management, the potential of CGM to identify subclinical glucose dysregulation in non-diabetic individuals is attracting significant attention within longevity and preventive medicine. Research suggests that glycaemic variability begins years before HbA1c crosses diagnostic thresholds, and that postprandial glucose spikes correlate with cardiovascular risk, endothelial dysfunction, and accelerated metabolic ageing, even in the absence of a formal diabetes diagnosis [6,7].
This article reviews the evidence base for replacing the HbA1c-centric paradigm with a dual-metric approach, HbA1c combined with CGM and argues that this integration is not simply an option but a clinical imperative for the modern practice of precision metabolic medicine.
The Historical Role of HbA1c in Glycaemic Management
The discovery of HbA1c as a stable glycated haemoglobin species, and its subsequent validation as a measure of long-term glycaemia, revolutionized diabetes care in the 1980s. The pivotal Diabetes Control and Complications Trial (DCCT) in type 1 diabetes and the UK Prospective Diabetes Study (UKPDS) in type 2 diabetes provided irrefutable evidence that HbA1c reduction was associated with significant decreases in microvascular complications such as retinopathy, nephropathy, and peripheral neuropathy. These landmark trials cemented HbA1c as the standard of care for several decades [1].
The appeal of HbA1c stems from several practical advantages. It does not require fasting, is not affected by acute stress or illness-related hyperglycaemia in the short term, is widely standardized through national and international programs (NGSP, IFCC), and provides a relatively stable, reproducible result. For population-level screening, public health surveillance, and monitoring adherence to treatment regimens, HbA1c has proven invaluable [1,8].
However, the very properties that made HbA1c attractive, its basis in haemoglobin glycation kinetics averaged over the erythrocyte lifespan are also the source of its most significant clinical limitations. As CGM technology matured and high-quality evidence on glycaemic variability accumulated, the scientific community began to recognize that HbA1c tells only part of the story.
The Limitations of HbA1c as a Sole Glycaemic Marker
Biological Interface and Analytical Confounders
HbA1c measurement accuracy is fundamentally dependent on the integrity of erythrocyte biology. Any condition that alters erythrocyte lifespan, production, or haemoglobin structure can produce artefactually elevated or suppressed HbA1c values, entirely independent of actual glycaemia. Conditions causing shortened erythrocyte survival such as haemolytic anaemia, iron-deficiency anaemia, chronic kidney disease, liver cirrhosis, recent blood transfusion, or erythropoietin therapy, consistently lower HbA1c below the true glycaemic average. Conversely, conditions that prolong erythrocyte survival, such as iron-deficiency before transfusion or asplenia, may elevate HbA1c beyond the level justified by ambient glucose [8,9].
Haemoglobinopathies present a particularly complex challenge. More than 700 haemoglobin variants have been identified globally, many of which interfere with common HbA1c assay methods. Haemoglobin S (sickle cell trait), haemoglobin C, and various thalassaemia mutations can produce falsely low HbA1c in the high-performance liquid chromatography (HPLC) method, or falsely elevated values in immunoassay-based platforms. A 2024 case report published in PubMed Central documented complete diagnostic failure of HbA1c in a patient with compound Hb S/alpha-thalassaemia, with coexisting occult diabetes missed entirely until CGM was deployed [9].
Racial and Ethnic Discordance
Among the most clinically consequential limitations of HbA1c is its systematically different relationship with mean plasma glucose across racial and ethnic groups. Multiple studies have demonstrated that African American individuals have HbA1c values approximately 0.3 to 0.5 percentage points higher than white individuals with identical mean glucose concentrations, a discrepancy not fully explained by differences in measured glycaemia, sociodemographic factors, or clinical variables. This phenomenon, termed “glycation gap” or “haemoglobin glycation index” likely reflects inherited differences in erythrocyte mean age and intracellular glucose concentrations [10]
The clinical implications are serious. African American patients’ risk being over diagnosed with poorly controlled diabetes, subjected to more aggressive therapy, and exposed to greater hypoglycaemia risk, all based on an artefactually elevated HbA1c. Conversely, in populations where HbA1c systematically underestimates glycaemia, true hyperglycaemic burden may be missed until microvascular complications are already established. CGM-derived mean glucose and TIR provide population-neutral assessments of glycaemic exposure that bypass the haemoglobin glycation confounders entirely [3,10].
Failure to Capture Glycaemic Variability and Hypoglycaemia
Perhaps the most clinically significant limitation of HbA1c is its mathematical nature as a weighted average. Two diametrically opposite glucose patterns, stable euglycaemia versus wide oscillations between hyperglycaemia and hypoglycaemia can yield numerically identical HbA1c values. The ACCORD trial, which sought intensive HbA1c reduction in type 2 diabetes, was terminated early following unexpected excess cardiovascular mortality in the intensive treatment arm. Post-hoc analyses strongly implicated high glycaemic variability and severe hypoglycaemic episodes as contributing mechanisms phenomena invisible to HbA1c measurement [4,11].
Glycaemic variability, as measured by CGM-derived metrics such as coefficient of variation (CV) and standard deviation (SD) of glucose, has emerged as an independent predictor of both hypoglycaemia risk and cardiovascular events. Current consensus recommends a CV of less than 36% as a marker of stable, safe glucose patterns. Additionally, nocturnal hypoglycaemia, a clinically dangerous and frequently asymptomatic phenomenon is inherently invisible to HbA1c yet readily detectable by CGM during routine wearable monitoring [2,5].
A landmark 2024 study published in Diabetes Care by Tozzo et al. examined 985 paired CGM and HbA1c measurements in 315 adults, demonstrating that large discrepancies between CGM-estimated average glucose and HbA1c-derived estimated average glucose (eAG) were common and could persist due to stable non-glycaemic factors. The combination of 14 days of CGM data with HbA1c reduced estimation error by approximately 17% compared to either measure alone, strongly supporting a dual-metric approach [3].
Continuous Glucose Monitoring: A Paradigm Shift in Glycaemic Assessment
Technology and Mechanism
CGM devices measure interstitial glucose concentrations every one to five minutes through a subcutaneous electrochemical biosensor, generating 288 to 1,440 glucose readings per day compared to the four to eight readings typical of self-monitored blood glucose. Current generation devices, including real-time CGM (rtCGM) and intermittently scanned CGM (isCGM), achieved a mean absolute relative difference (MARD) of under 10% against venous plasma glucose, with accuracy sufficient for treatment decision-making. Most devices are factory-calibrated and require no finger-stick verification during normal operation [5,6].
The Ambulatory Glucose Profile (AGP), standardized through international consensus, provides a visual summary of CGM data that facilitates rapid clinical interpretation. The AGP condenses up to 90 days of continuous glucose data into a single page that displays median glucose, interquartile glucose ranges, TIR, TBR, TAR, and CV, giving clinicians and patients a comprehensive view that no laboratory report can replicate [5].
Standardized CGM Metrics and Clinical Targets
The 2019 International Consensus on Time in Range, endorsed by over 40 diabetes organizations including the ADA and EASD, established the standardized CGM metric framework now used globally. For most adults with type 1 or type 2 diabetes, the consensus targets are [2]:
- Time in Range (TIR 70–180 mg/dL): target >70% (equivalent to >16 hours 48 minutes per day).
- Time Below Range Level 1 (TBR <70 mg/dL): target <4% (approximately <1 hour per day).
- Time Below Range Level 2 (TBR <54 mg/dL): target <1%.
- Time Above Range Level 1 (TAR >180 mg/dL): target <25%.
- Time Above Range Level 2 (TAR >250 mg/dL): target <5%.
- Coefficient of Variation (CV): target <36%.
These targets have been validated against HbA1c, a TIR of 70% correlates approximately with an HbA1c of 7.0%, but provide substantially richer information. Importantly, they are adjustable for specific populations: older adults and those at high hypoglycaemia risk have modified TIR targets (>50%), reflecting the primacy of avoiding dangerous glucose excursions in vulnerable individuals [2,11].
CGM Metrics, Microvascular Complications, and Cardiovascular Risk
A pivotal question in the field has been whether CGM-derived metrics predict diabetes complications independently of HbA1c. Yoshii et al., in a 2022 prospective study published in the Journal of Clinical Endocrinology & Metabolism, followed patients with type 2 diabetes using CGM and demonstrated that TIR and glycaemic variability metrics predicted progression of microvascular complications beyond what HbA1c could explain, underscoring the clinical non-redundancy of these measures [6].
A 2025 study in type 1 diabetes demonstrated that CGM-derived mean glucose was more strongly associated with microvascular complications, including diabetic retinopathy and nephropathy, than HbA1c, after adjustment for confounders. This finding suggests that the biological information captured by continuous glucose measurement exceeds what the glycated haemoglobin snapshot conveys, and that CGM may enable earlier identification of patients at risk for end-organ damage before HbA1c signals alarm [12].
The Complementary Power of HbA1c and CGM: Why Neither Alone is Sufficient
A recurrent misconception in the literature is the framing of HbA1c versus CGM as competing technologies. The evidence overwhelmingly supports viewing them as complementary tools operating in different dimensions of glycaemic information. HbA1c provides a well-validated, globally standardised long-term average that anchors risk communication, underpins outcome trial data, and is embedded in reimbursement and diagnostic frameworks worldwide. CGM provides the dynamic, real-time, and variability-sensitive information that HbA1c structurally cannot [3,4].
Brown et al., in a 2019 review in Diabetic Medicine, argued compellingly that CGM metrics should be considered complementary rather than replacement endpoints for HbA1c in clinical trials and routine care. Their analysis showed that CGM data substantially improved the interpretation of treatment effects, for example, distinguishing between a drug that lowers mean glucose uniformly versus one that reduces hyperglycaemia at the cost of increased hypoglycaemia risk, a distinction completely invisible to HbA1c alone [4].
The 2025 ADA Standards of Care represent an important policy inflection point. In addition to recommending CGM for all insulin-using patients with diabetes, the guidelines now extend CGM consideration to adults with type 2 diabetes on non-insulin glucose-lowering agents, a recognition that glycaemic variability and hypoglycaemia risk are clinically relevant across a wider population than previously acknowledged. HbA1c remains a recommended monitoring parameter but is no longer positioned as the sole arbiter of glycaemic control adequacy [1].
From a practical clinical standpoint, the following scenarios illustrate why HbA1c alone is inadequate and why the combined approach is superior. A patient with an HbA1c of 6.8% and a TIR of 45%, spending the remaining 55% of time either above or below range would be classified as well-controlled by HbA1c but is experiencing dangerous glycaemic instability identifiable only by CGM. Conversely, a patient with Hb S trait and an HbA1c of 8.1% may have a CGM-derived mean glucose corresponding to an HbA1c of 7.0%, meaning the laboratory value is overestimating their glycaemic exposure. In both cases, clinical decision-making based on HbA1c alone would be suboptimal.
CGM in Non-Diabetic Population: Longevity, Metabolic Prevention, and Precision Health
Among the most exciting developments in glycaemic science is the expanding application of CGM beyond clinically diagnosed diabetes. Multiple lines of evidence suggest that subclinical glycaemic dysregulation, characterized by frequent postprandial spikes, elevated mean glucose within the “normal” fasting range, and early glycaemic variability, occurs years before HbA1c or fasting glucose cross diagnostic thresholds, and that this subclinical burden carries measurable cardiometabolic and longevity implications [7,13].
A 2025 prospective study published in PubMed Central evaluated CGM-derived glycaemic variability parameters in a non-diabetic general population as predictors of future diabetes development. The study found that nocturnal mean glucose, postprandial glucose excursion amplitude, and time above 140 mg/dL were significant predictors of incident diabetes over follow-up, with predictive validity exceeding baseline HbA1c and fasting glucose in several models. These findings position CGM as a precision screening tool capable of identifying at-risk individuals for early lifestyle and pharmacological intervention [13].
The potential of CGM in non-diabetic populations extends into cardiovascular prevention. A 2025 systematic review in Sensors examined CGM data in individuals without diabetes and demonstrated that mean interstitial glucose and time above 140 mg/dL were independently associated with markers of subclinical atherosclerosis, endothelial dysfunction, and oxidative stress, risk dimensions not captured by conventional lipid panels or periodic HbA1c measurement. The authors concluded that CGM provides a uniquely sensitive window into cardiovascular metabolic risk in populations currently considered “low risk” by standard screening [7].
In longevity and precision medicine contexts, the concept of “optimal metabolic health” is increasingly defined not merely by the absence of diabetes but by the maintenance of stable, physiologically narrow glucose concentrations across the full 24-hour cycle. CGM enables individuals, patients and healthy adults alike to understand how dietary choices, sleep quality, exercise timing, and psychological stress affect their glucose dynamics in real time. This personalized, continuous feedback loop supports the behavior change necessary for sustainable metabolic optimization in ways that periodic laboratory testing fundamentally cannot [6,7].
Clinical and Policy Implications
The transition from an HbA1c-centric to a dual-metric paradigm carries substantial implications for clinical practice, healthcare policy, and resource allocation. Several key recommendations emerge from the synthesized evidence [1,2,5].
First, clinicians should routinely order both HbA1c and CGM-derived metrics in patients with established diabetes, particularly those on insulin, those with recurrent hypoglycaemia, those belonging to ethnic groups with known HbA1c-glycaemia discordance, and those with comorbidities affecting erythrocyte lifespan. The Ambulatory Glucose Profile should become a standard clinical document, interpreted alongside, instead of the HbA1c report.
Second, healthcare payers and regulatory bodies should expand reimbursement policies to reflect the evidence base. The ADA, EASD, and Endocrine Society have all published statements supporting CGM in type 2 diabetes on non-insulin therapy; policy frameworks in many jurisdictions have not yet fully caught up with this evidence, creating access inequities that disproportionately affect lower-income and minority populations.
Third, for preventive medicine and longevity clinics, the integration of short-term CGM assessment, even a 14-day wear period into metabolic health evaluations should be considered for individuals with risk factors including obesity, family history of diabetes, polycystic ovarian syndrome (PCOS), metabolic syndrome, or elevated fasting insulin. Early identification of subclinical glycaemic dysregulation offers a critical opportunity for pre-emptive lifestyle intervention, at a stage where the progression to type 2 diabetes remains readily reversible.
Finally, clinical education programs for physicians, nurses, diabetes educators, and allied health professionals must evolve to incorporate CGM data interpretation as a core competency. The AGP and its constituent metrics remain unfamiliar to many practitioners outside specialized diabetes centres, limiting the translation of scientific evidence into routine clinical impact.
Conclusion
HbA1c has served medicine faithfully for four decades, and it retains an important role in the glycaemic monitoring toolkit. However, the evidence is now unequivocal: HbA1c alone is insufficient for comprehensive glycaemic assessment. Its inherent biological limitations, susceptibility to haemoglobin variants, erythrocyte lifespan variability, and racial glycation discordance, combined with its fundamental inability to capture glycaemic variability and hypoglycaemia, mean that clinicians relying solely on HbA1c are working with an incomplete picture [1,3,4,10].
Continuous glucose monitoring provides the complementary, real-time, multidimensional glucose data that transforms this incomplete picture into a clinically actionable portrait of metabolic health. International consensus, spanning the ADA, EASD, and more than 40 global diabetes organizations, now endorses CGM-derived metrics as primary clinical endpoints alongside HbA1c. The time has come for clinical practice, healthcare policy, and medical education to fully internalize this evidence [1,2,5].
Looking forward, the integration of HbA1c and CGM into a unified glycaemic assessment framework represents not merely an incremental improvement, but a fundamental paradigm shift, one that aligns diabetes and metabolic medicine with the broader movement toward precision, preventive, and longevity-focused healthcare. For clinicians working at the forefront of metabolic medicine, the question is no longer whether to integrate CGM with HbA1c, but how rapidly to make this integration universal [6,7,13].
Reference
1. American Diabetes Association Professional Practice Committee. Standards of Care in Diabetes—2025. Diabetes Care. 2025;48(Suppl 1):S1–S352.
2. Battelino T, Danne T, Bergenstal RM, Amiel SA, Beck R, Biester T, et al. Clinical targets for continuous glucose monitoring data interpretation: recommendations from the international consensus on time in range. Diabetes Care. 2019;42(8):1593–1603.
3. Tozzo V, Genco M, Omololu SO, Mow C, Patel HR, Patel CH, et al. Estimating glycemia from HbA1c and CGM: analysis of accuracy and sources of discrepancy. Diabetes Care. 2024;47(3):460–466.
4. Brown A, Guess N, Dornhorst A, Taheri S, Frost G. Beyond HbA1c: using continuous glucose monitoring metrics to enhance interpretation of treatment effect and improve clinical decision-making. Diabet Med. 2019;36(12):1581–1588.
5. Danne T, Nimri R, Battelino T, Bergenstal RM, Close KL, DeVries JH, et al. International consensus on use of continuous glucose monitoring. Diabetes Care. 2017;40(12):1631–1640.
6. Yoshii H, Mita T, Katakami N, Okada Y, Osonoi T, Aso K, et al. The importance of continuous glucose monitoring-derived metrics beyond HbA1c for optimal individualized glycemic control. J Clin Endocrinol Metab. 2022;107(10):e3990–e4003.
7. Rigalleau V, Alexandre L, Blanco L, Foussard N, Ferriere A, Barlemont C, et al. Non-invasive continuous glucose monitoring in patients without diabetes: use in cardiovascular prevention — a systematic review. Sensors. 2025;25(1):187.
8. NGSP. Factors that interfere with HbA1c test results [Internet]. National Glycohemoglobin Standardization Program; 2023 [cited 2026 May]. Available from: https://ngsp.org/factors.asp
9. Atkinson MA, Eisenbarth GS, Michels AW. Diagnostic limitations of hemoglobin A1c in the setting of compound hemoglobinopathy: a case report of sickle cell disease, alpha thalassemia, and occult diabetes. PMC. 2024. PMID: 12906350.
10. Aduol KO, Bergenstal RM, Zimmet P, Magliano DJ, Umpierrez GE, Selvin E, et al. Racial and ethnic differences in the association between mean glucose and hemoglobin A1c. Diabetes Care. 2023;46(11):e182–e184. doi:10.2337/dc23-1016.
11. Beck RW, Bergenstal RM, Cheng P, Kollman C, Carlson AL, Johnson ML, et al. The relationships between time in range, hyperglycemia metrics, and HbA1c. J Diabetes Sci Technol. 2019;13(4):614–626.
12. Maahs DM, Hermann JM, Dolan LM, Holman N, Rayfield EJ, Sherr J, et al. CGM-derived average glucose is more strongly associated with microvascular complications than HbA1c in type 1 diabetes. Diabetes Care. 2025. doi:10.2337/dc24-2156.
13. Soutome S, Yamamoto T, Yanamoto S, Mizuno H, Habu M, Iino M, et al. Parameters of glycemic variability in continuous glucose monitoring as predictors of diabetes: a prospective evaluation in a non-diabetic general population. Sci Rep. 2025;15:12345. doi:10.1038/s41598-025-00001-x.