The sensitivity and specificity of HbA1c are central to understanding its value and limitations as a diagnostic test for type 2 diabetes. HbA1c, or glycated haemoglobin, reflects average blood glucose over the preceding two to three months and is widely used across the NHS for both diagnosis and monitoring. At the WHO-recommended threshold of 48 mmol/mol, HbA1c offers high specificity but moderate sensitivity, meaning it reliably confirms diabetes when positive yet may miss a proportion of true cases. Clinicians must understand these diagnostic properties alongside the clinical, biological, and population factors that can affect result accuracy.

What HbA1c Measures and How It Is Used in the NHS

HbA1c measures average blood glucose over two to three months and is used by the NHS both to diagnose type 2 diabetes in adults and to monitor glycaemic control in people already living with diabetes.

HbA1c, or glycated haemoglobin, reflects the average blood glucose concentration over the preceding two to three months. When glucose circulates in the bloodstream, it binds irreversibly to haemoglobin within red blood cells through a process called glycation. Because red blood cells have a lifespan of approximately 120 days, the proportion of haemoglobin that has become glycated provides a reliable retrospective indicator of glycaemic control — unlike a fasting plasma glucose measurement, which captures only a single moment in time.

Within the NHS, HbA1c serves two distinct clinical purposes. First, it is used as a diagnostic tool to identify type 2 diabetes in adults in appropriate circumstances. Second, it is used as a monitoring tool for people already living with diabetes of any type, helping clinicians and patients assess how well blood glucose is being managed over time and guiding treatment decisions. Results are expressed in millimoles per mole (mmol/mol) in the UK, following the IFCC (International Federation of Clinical Chemistry) standardisation adopted nationally.

Importantly, HbA1c should not be used to diagnose suspected type 1 diabetes (NICE NG17). Adults with symptoms suggestive of type 1 diabetes require urgent same-day assessment using plasma or capillary blood glucose, not HbA1c.

The test must be performed on a venous blood sample analysed in an accredited laboratory using an IFCC-aligned method. Point-of-care (capillary) HbA1c devices must not be used for diagnostic purposes, as they do not meet the required analytical standards. The test does not require the patient to fast beforehand, which makes it more convenient and reproducible than oral glucose tolerance testing (OGTT).

NICE guidance (NG28) supports HbA1c as a first-line diagnostic test in adults who are not acutely unwell and do not have conditions known to interfere with the result. HbA1c is not appropriate for diagnosis in children and young people, during pregnancy, in acutely unwell patients, or in individuals with conditions affecting red cell survival or haemoglobin structure (see the section on factors affecting accuracy below). Understanding the sensitivity and specificity of HbA1c is essential for interpreting results correctly and avoiding both over- and under-diagnosis.

Sensitivity and Specificity of HbA1c for Diagnosing Diabetes

At the 48 mmol/mol threshold, HbA1c has high specificity (~95–99%) but moderate sensitivity (~44–66%), so a positive result strongly indicates diabetes, but a negative result does not reliably exclude it.

The sensitivity of a diagnostic test refers to its ability to correctly identify individuals who truly have the condition — in other words, the proportion of people with diabetes who will receive a positive result. The specificity refers to its ability to correctly identify those who do not have the condition, thereby avoiding false positives. Both metrics are critical when evaluating HbA1c as a diagnostic instrument.

Research comparing HbA1c against the oral glucose tolerance test (OGTT) — historically considered the reference standard for diabetes diagnosis — has consistently shown that HbA1c at the WHO-recommended diagnostic threshold of 48 mmol/mol (6.5%) has relatively high specificity (approximately 95–99%) but moderate sensitivity (approximately 44–66%). This means the test is very good at confirming diabetes when the result is positive, but it may miss a meaningful proportion of individuals who would be identified as having diabetes by OGTT. The precise figures vary across studies depending on the population studied, ethnicity, and the reference standard used (WHO, 2011; UK consensus statement/PHE, 2012).

It is also important to recognise that the positive and negative predictive values of HbA1c depend on the pre-test probability of diabetes in the individual being tested. In a high-risk population, a positive result carries greater certainty; in a low-risk population, even a high-specificity test will generate a higher proportion of false positives relative to true positives. Clinical context should therefore always inform interpretation.

The clinical implication of the sensitivity–specificity asymmetry is important:

• A result of ≥48 mmol/mol on two separate occasions (or once alongside symptoms) is highly likely to represent true diabetes — false positives are uncommon.

• A result below 48 mmol/mol does not definitively exclude diabetes, particularly in individuals with strong risk factors or symptoms suggestive of hyperglycaemia.

For the non-diabetic hyperglycaemia range (42–47 mmol/mol), sensitivity and specificity vary further, and clinicians should interpret borderline results in the context of the full clinical picture. Where clinical suspicion remains high despite a sub-threshold HbA1c, alternative testing such as fasting plasma glucose or OGTT should be considered.

Factors That Affect HbA1c Accuracy and Reliability

Conditions including haemolytic anaemia, haemoglobin variants, pregnancy, and advanced chronic kidney disease can falsely lower or elevate HbA1c, requiring alternative diagnostic methods such as fasting plasma glucose or OGTT.

HbA1c is a robust test under most circumstances, but a range of clinical and biological factors can render the result unreliable — either falsely elevated or falsely lowered — independent of true glycaemic status. Recognising these factors is essential for patient safety and accurate diagnosis. Where any of the following conditions are present or suspected, NICE and NHS guidance recommend using alternative diagnostic methods such as fasting plasma glucose or OGTT rather than relying on HbA1c alone (NICE NG28; UK consensus statement/PHE, 2012).

Conditions that may falsely lower HbA1c:

• Haemolytic anaemia — increased red cell turnover reduces the time available for glycation, lowering HbA1c disproportionately to actual glucose levels.

• Iron deficiency anaemia treated with iron — repletion increases red cell production, diluting glycated cells.

• Haemoglobin variants (e.g., HbS, HbC, HbE) — common in individuals of African, South Asian, and Mediterranean heritage; certain variants interfere with laboratory assay methods. The degree of interference is assay-dependent; laboratories may use variant-insensitive methods and will flag potential interference. Clinicians should check local laboratory notes and consider alternative tests if interference is suspected (RCPath/UK NEQAS guidance).

• Pregnancy — physiological haemodilution and altered red cell turnover make HbA1c unreliable, particularly in the second and third trimesters.

• Recent blood transfusion or significant blood loss — introduces donor red cells or alters red cell mass, distorting the result.

• Erythropoietin-stimulating agent (ESA) therapy — accelerates red cell turnover, reducing HbA1c independently of glycaemia.

• Advanced chronic kidney disease (CKD) — in advanced CKD (particularly eGFR <30 ml/min/1.73 m²) and in patients receiving ESA therapy, reduced red cell survival frequently causes HbA1c to be falsely low, making it an unreliable indicator of glycaemic control. Plasma glucose or OGTT should be used in these patients.

Conditions that may falsely elevate HbA1c:

• Iron deficiency anaemia (untreated) — reduced red cell turnover prolongs exposure to glucose, increasing glycation.

• Vitamin B12 or folate deficiency — similar mechanism of reduced erythropoiesis.

• Splenectomy — prolonged red cell lifespan increases glycation time.

• Severe liver disease — may affect red cell turnover and haemoglobin metabolism.

Note on chronic kidney disease: The effect of CKD on HbA1c is complex and bidirectional. Whilst reduced erythropoiesis in early CKD may cause mild elevation, advanced CKD and ESA use more commonly result in falsely low values. Overall reliability is significantly reduced; plasma glucose measurements are preferred.

Clinicians should also be aware that ethnicity-related differences in HbA1c levels have been documented, with some studies suggesting that HbA1c may be systematically higher in people of Black African or South Asian heritage at equivalent glucose levels. However, no ethnicity-specific HbA1c diagnostic thresholds are recommended in the UK. Where there is a discordance between HbA1c and clinical suspicion, confirmatory glucose testing should be considered.

NICE and WHO Guidance on HbA1c Diagnostic Thresholds

NICE and WHO recommend a threshold of 48 mmol/mol (6.5%) for diagnosing type 2 diabetes in adults, requiring two positive results in asymptomatic individuals or one result alongside classic hyperglycaemic symptoms.

The World Health Organisation (WHO) formally endorsed HbA1c as a diagnostic test for type 2 diabetes in 2011, recommending a threshold of 48 mmol/mol (6.5%) as the cut-off for diagnosis. This recommendation was adopted by NICE and is embedded in NHS clinical practice across England, Wales, Scotland, and Northern Ireland. The 48 mmol/mol threshold was selected primarily because it corresponds to the level at which the risk of diabetic retinopathy increases significantly — high specificity is a consequence of this evidence-based cut-off, rather than the primary design intent (WHO, 2011; NICE NG28).

According to NICE guidance (NG28 and associated updates), the following diagnostic criteria apply in adults:

• HbA1c ≥48 mmol/mol on two separate occasions confirms a diagnosis of type 2 diabetes in asymptomatic individuals.

• A single HbA1c ≥48 mmol/mol is sufficient for diagnosis when the patient has classic symptoms of hyperglycaemia (e.g., polyuria, polydipsia, unexplained weight loss).

• HbA1c of 42–47 mmol/mol indicates non-diabetic hyperglycaemia (NDH) (sometimes referred to as prediabetes), warranting lifestyle intervention and annual monitoring.

• HbA1c is not recommended for diagnosing type 1 diabetes, gestational diabetes, or diabetes in acutely unwell patients, where plasma glucose measurements remain the standard.

• Diagnosis must use a venous blood sample analysed in an accredited laboratory with an IFCC-traceable method. Point-of-care HbA1c devices are not acceptable for diagnostic purposes.

HbA1c cannot be used diagnostically in children and young people under current NICE guidance; clinical assessment and plasma glucose testing are preferred in paediatric populations.

Urgent same-day referral: Adults presenting with symptoms suggestive of type 1 diabetes (e.g., rapid onset of symptoms, significant weight loss, ketosis) or with features of a hyperglycaemic emergency such as diabetic ketoacidosis (DKA) or hyperosmolar hyperglycaemic state (HHS) must not have diagnosis delayed pending HbA1c. These individuals require urgent same-day assessment, diagnosis using plasma or capillary blood glucose, and immediate referral to a specialist diabetes team (NICE NG17).

From a patient safety perspective, individuals with an HbA1c in the non-diabetic hyperglycaemia range should be referred to the NHS Diabetes Prevention Programme (NDPP) where eligible. Those with symptoms of diabetes alongside a borderline or unexpectedly low HbA1c should be advised to contact their GP promptly for further evaluation, as the limitations of the test mean that a normal result does not always exclude the condition. Clinicians should document any factors that may affect HbA1c reliability and record the rationale for choosing alternative diagnostic approaches where applicable.

Frequently Asked Questions