Diabetes is diagnosed using biochemical tests that demonstrate abnormal glucose regulation. The principal tests for non-pregnant individuals are HbA1c, fasting plasma glucose, the two-hour plasma glucose concentration during a 75-g oral glucose tolerance test and, in people with classic hyperglycaemic symptoms or a hyperglycaemic crisis, random plasma glucose. HbA1c measures the proportion of haemoglobin that has glucose attached to it and therefore reflects average glycaemic exposure over approximately the preceding two to three months.
Fasting plasma glucose measures glucose after at least eight hours without caloric intake. The oral glucose tolerance test evaluates the body's response to a defined glucose load. Random plasma glucose can be diagnostic when it reaches the relevant threshold in an appropriate symptomatic context. The exact testing pathway depends on symptoms, pregnancy status, clinical context and local guidelines.
For non-pregnant individuals, commonly used diagnostic thresholds are HbA1c of at least 6.5% (48 mmol/mol), fasting plasma glucose of at least 7.0 mmol/L, or two-hour plasma glucose of at least 11.1 mmol/L during a 75-g oral glucose tolerance test. A random plasma glucose of at least 11.1 mmol/L can support diagnosis when accompanied by classic symptoms of hyperglycaemia or hyperglycaemic crisis. If there is no unequivocal hyperglycaemia, the diagnosis generally requires confirmation with a second abnormal test or repeat testing.
These thresholds are diagnostic criteria rather than descriptions of every person's ideal glucose level. A result just below a threshold does not mean that metabolic risk suddenly disappears, because glucose-related risk exists on a continuum. ADA criteria define prediabetes as HbA1c 5.7–6.4%, fasting glucose 5.6–6.9 mmol/L or two-hour glucose 7.8–11.0 mmol/L.
HbA1c is convenient because it does not require fasting and represents longer-term average glucose exposure. However, HbA1c does not measure glucose directly and can be misleading when red blood-cell lifespan or haemoglobin characteristics are abnormal. Conditions involving altered red-cell turnover, certain haemoglobin variants, pregnancy and some other clinical circumstances can change the relationship between HbA1c and actual glycaemia. The ADA therefore recommends using plasma glucose criteria when the relationship between HbA1c and glucose is likely to be altered.
HbA1c can also fail to capture rapid glucose changes because two people with the same average glucose may have very different daily patterns. Consequently, HbA1c is valuable for diagnosis and monitoring but should be interpreted in clinical context. Where a laboratory glucose result and HbA1c appear substantially discordant, clinicians may investigate whether one measurement is affected by an underlying condition or analytical limitation.
Fasting plasma glucose requires an appropriate fasting period, generally at least eight hours without caloric intake. A result of 7.0 mmol/L or higher meets the diagnostic threshold used by major guidelines for non-pregnant adults, although confirmation is normally required if unequivocal hyperglycaemia is absent. The oral glucose tolerance test is more demanding but provides information about how effectively the body handles a defined glucose challenge.
After an overnight fast, the person drinks a solution containing 75 g of glucose and blood glucose is measured at baseline and two hours later. A two-hour value of 11.1 mmol/L or higher meets the diagnostic criterion for diabetes. Intermediate results indicate impaired glucose tolerance under commonly used criteria. Because fasting glucose and two-hour glucose measure different aspects of glucose physiology, they do not identify exactly the same populations, so one test can be abnormal while another remains below the diagnostic threshold.
Random plasma glucose is useful when someone presents with classic symptoms because it avoids delaying diagnosis while waiting for a fasting measurement. A random glucose of at least 11.1 mmol/L together with symptoms such as excessive urination, excessive thirst and unexplained weight loss can establish diabetes in the appropriate clinical setting. A high random glucose without symptoms is not necessarily sufficient on its own and generally requires confirmation.
Very high glucose associated with vomiting, dehydration, abnormal breathing, altered consciousness or other signs of metabolic crisis requires urgent assessment rather than routine outpatient confirmation. Clinicians may also test ketones, electrolytes, renal function and acid-base status when DKA is suspected. Thus, diabetes diagnosis and emergency assessment are related but distinct processes: the first establishes chronic metabolic disease, while the second determines whether acute metabolic decompensation is occurring.
Diagnosis during pregnancy requires special consideration because pregnancy changes glucose physiology and because gestational diabetes has pregnancy-specific thresholds. The ADA recommends screening for gestational diabetes at 24–28 weeks in pregnant individuals who have not previously been diagnosed with diabetes or high-risk abnormal glucose metabolism. A one-step strategy uses a 75-g oral glucose tolerance test with fasting, one-hour and two-hour measurements, with diagnostic thresholds of 5.1, 10.0 and 8.5 mmol/L respectively under the cited criteria.
A two-step strategy uses an initial 50-g glucose challenge followed, when indicated, by a fasting 100-g oral glucose tolerance test. Different organisations and jurisdictions use somewhat different thresholds and testing strategies, so pregnancy results should be interpreted according to the guideline used by the maternity service.
Diagnosis should be distinguished from screening. Screening identifies people who should undergo diagnostic evaluation, often because they have risk factors or belong to a population in which diabetes is more prevalent. Diagnostic testing determines whether biochemical criteria for diabetes are actually met. Prediabetes is an intermediate category rather than simply “early diabetes” in every case. It signals abnormal glucose regulation and increased future risk but does not automatically progress to diabetes.
Screening can involve HbA1c, fasting glucose or other appropriate tests, while continuous glucose monitoring is not currently considered a standard diagnostic screening test for diabetes in the ADA guidance. Home glucose meters are useful for monitoring but are not generally interchangeable with laboratory diagnostic testing because measurement methods and accuracy requirements differ.
Once diabetes is diagnosed, additional testing is needed to classify the condition and establish a baseline for management. Clinicians may assess autoantibodies when type 1 diabetes is suspected, C-peptide when endogenous insulin production is uncertain, lipid levels, kidney function, blood pressure, urine albumin, eye health and foot health. The purpose is not merely to label the disease but to determine its mechanism, severity, complications and appropriate treatment.
In children or adults with atypical presentation, strong multigenerational family history or unusual treatment response, genetic testing may sometimes identify monogenic diabetes. Classification is clinically important because insulin is essential for type 1 diabetes, whereas type 2 diabetes has a much broader range of pharmacological and non-pharmacological treatment options. Diagnosis therefore consists of biochemical confirmation followed by clinical classification rather than a single isolated blood test producing the entire answer.
