The gold standard for investigating and quantifying insulin resistance is the "hyperinsulinemic euglycemic clamp," so-called because it measures the amount of glucose necessary to compensate for an increased insulin level without causing hypoglycemia. It is a type of glucose clamp technique. The test rarely is performed in clinical care, but is used in medical research, for example, to assess the effects of different medications. The rate of glucose infusion commonly is referred to in diabetes literature as the GINF value.

The procedure takes about two hours. Through a peripheral vein, insulin is infused at 10–120 mU per m per minute. In order to compensate for the insulin infusion, glucose 20% is infused to maintain blood sugar levels between 5 and 5.5 mmol/L. The rate of glucose infusion is determined by checking the blood sugar levels every five to ten minutes.

The rate of glucose infusion during the last thirty minutes of the test determines insulin sensitivity. If high levels (7.5 mg/min or higher) are required, the patient is insulin-sensitive. Very low levels (4.0 mg/min or lower) indicate that the body is resistant to insulin action. Levels between 4.0 and 7.5 mg/min are not definitive, and suggest "impaired glucose tolerance," an early sign of insulin resistance.

This basic technique may be enhanced significantly by the use of glucose tracers. Glucose may be labeled with either stable or radioactive atoms. Commonly used tracers are 3-H glucose (radioactive), 6,6 H-glucose (stable) and 1-C Glucose (stable). Prior to beginning the hyperinsulinemic period, a 3h tracer infusion enables one to determine the basal rate of glucose production. During the clamp, the plasma tracer concentrations enable the calculation of whole-body insulin-stimulated glucose metabolism, as well as the production of glucose by the body (i.e., endogenous glucose production).

Another measure of insulin resistance is the modified insulin suppression test developed by Gerald Reaven at Stanford University. The test correlates well with the euglycemic clamp, with less operator-dependent error. This test has been used to advance the large body of research relating to the metabolic syndrome.

Patients initially receive 25 μg of octreotide (Sandostatin) in 5 mL of normal saline over 3 to 5 minutes via intravenous infusion (IV) as an initial bolus, and then, are infused continuously with an intravenous infusion of somatostatin (0.27 μg/m/min) to suppress endogenous insulin and glucose secretion. Next, insulin and 20% glucose are infused at rates of 32 and 267 mg/m/min, respectively. Blood glucose is checked at zero, 30, 60, 90, and 120 minutes, and thereafter, every 10 minutes for the last half-hour of the test. These last four values are averaged to determine the steady-state plasma glucose level (SSPG). Subjects with an SSPG greater than 150 mg/dL are considered to be insulin-resistant.

When the cause of hypoglycemia is not obvious, the most valuable diagnostic information is obtained from a blood sample (a "critical specimen") drawn during the hypoglycemia. Detectable amounts of insulin are abnormal and indicate that hyperinsulinism is likely to be the cause. Other aspects of the person's metabolic state, especially low levels of free fatty acids, beta-hydroxybutyrate and ketones, and either high or low levels of C-peptide and proinsulin can provide confirmation.

Clinical features and circumstances can provide other indirect evidence of hyperinsulinism. For instance, babies with neonatal hyperinsulinism are often large for gestational age and may have other features such as enlarged heart and liver. Knowing that someone takes insulin or oral hypoglycemic agents for diabetes obviously makes insulin excess the presumptive cause of any hypoglycemia.

Most sulfonylureas and aspirin can be detected on a blood or urine drug screen tests, but insulin cannot. Endogenous and exogenous insulin can be distinguished by the presence or absence of C-peptide, a by-product of endogenous insulin secretion which is not present in pharmaceutical insulin. Some of the newer analog insulins are not measured by the usual insulin level assays.

In terms of the investigation of congenital hyperinsulinism, valuable diagnostic information is obtained from a blood sample drawn during hypoglycemia, detectable amounts of insulin during hypoglycemia are abnormal and indicate that hyperinsulinism is likely to be the cause. Inappropriately low levels of free fatty acids and ketones provide additional evidence of insulin excess. An additional piece of evidence indicating hyperinsulinism is a usually high requirement for intravenous glucose to maintain adequate glucose levels, the minimum glucose required to maintain a plasma glucose above 70 mg/dl. A GIR above 8 mg/kg/minute in infancy suggests hyperinsulinism. A third form of evidence suggesting hyperinsulinism is a rise of the glucose level after injection of glucagon at the time of the low glucose.

Diagnostic efforts then shift to determining the type- elevated ammonia levels or abnormal organic acids can indicate specific, rare types. Intrauterine growth retardation and other perinatal problems raise the possibility of transience, while large birthweight suggests one of the more persistent conditions. Genetic screening is now available within a useful time frame for some of the specific conditions.It is worthwhile to identify the minority of severe cases with focal forms of hyperinsulinism because these can be completely cured by partial pancreatectomy. A variety of pre-operative diagnostic procedures have been investigated but none has been established as infallibly reliable. Positron emission tomography is becoming the most useful imaging technique.

There are differenct types of congenital hyperinsulinism as "diffuse and focal" indicated below:

Opinions differ about optimal screening and diagnostic measures, in part due to differences in population risks, cost-effectiveness considerations, and lack of an evidence base to support large national screening programs. The most elaborate regimen entails a random blood glucose test during a booking visit, a screening glucose challenge test around 24–28 weeks' gestation, followed by an OGTT if the tests are outside normal limits. If there is a high suspicion, a woman may be tested earlier.

In the United States, most obstetricians prefer universal screening with a screening glucose challenge test. In the United Kingdom, obstetric units often rely on risk factors and a random blood glucose test. The American Diabetes Association and the Society of Obstetricians and Gynaecologists of Canada recommend routine screening unless the woman is low risk (this means the woman must be younger than 25 years and have a body mass index less than 27, with no personal, ethnic or family risk factors) The Canadian Diabetes Association and the American College of Obstetricians and Gynecologists recommend universal screening. The U.S. Preventive Services Task Force found there is insufficient evidence to recommend for or against routine screening.

Some pregnant women and careproviders choose to forgo routine screening due to the absence of risk factors, however this is not advised due to the large proportion of women who develop gestational diabetes despite having no risk factors present and the dangers to the mother and baby if gestational diabetes remains untreated.

Women with GDM may have high glucose levels in their urine (glucosuria). Although dipstick testing is widely practiced, it performs poorly, and discontinuing routine dipstick testing has not been shown to cause underdiagnosis where universal screening is performed. Increased glomerular filtration rates during pregnancy contribute to some 50% of women having glucose in their urine on dipstick tests at some point during their pregnancy. The sensitivity of glucosuria for GDM in the first 2 trimesters is only around 10% and the positive predictive value is around 20%.

Children's blood sugar levels are often slightly lower than adults'. Overnight fasting glucose levels are below 70 mg/dL (3.9 mM) in 5% of healthy adults, but up to 5% of children can be below 60 mg/dL (3.3 mM) in the morning fasting state. As the duration of fasting is extended, a higher percentage of infants and children will have mildly low plasma glucose levels, usually without symptoms. The normal range of newborn blood sugars continues to be debated. It has been proposed that newborn brains are able to use alternate fuels when glucose levels are low more readily than adults. Experts continue to debate the significance and risk of such levels, though the trend has been to recommend maintenance of glucose levels above 60–70 mg/dL the first day after birth.

Diabetic hypoglycemia represents a special case with respect to the relationship of measured glucose and hypoglycemic symptoms for several reasons. First, although home glucose meter readings are often misleading, the probability that a low reading, whether accompanied by symptoms or not, represents real hypoglycemia is much higher in a person who takes insulin than in someone who does not.

The diagnosis is based on a combination of typical clinical features and exclusion by a pediatric endocrinologist of other causes of "hypoglycemia with ketosis," especially growth hormone deficiency, hypopituitarism, adrenal insufficiency, and identifiable inborn errors of metabolism such as organic acidoses.

The most useful diagnostic tests include measurement of insulin, growth hormone, cortisol, and lactic acid at the time of the hypoglycemia. Plasma acylcarnitine levels and urine organic acids exclude some of the important metabolic diseases. When the episodes are recurrent or severe, the definitive test is a hospitalization for a supervised diagnostic fast. This usually demonstrates "accelerated fasting"—a shorter time until the glucose begins to fall, but normal metabolic and counterregulatory responses as the glucose falls. As the glucose reaches hypoglycemic levels, the insulin is undetectable, counterregulatory hormones, fatty acids, and ketones are high, and glucagon injection elicits no rise of glucose.

Acute hypoglycemia is reversed by raising the blood glucose. Glucagon should be injected intramuscularly or intravenously, or dextrose can be infused intravenously to raise the blood glucose. Oral administration of glucose can worsen the outcome, as more insulin is eventually produced. Most people recover fully even from severe hypoglycemia after the blood glucose is restored to normal. Recovery time varies from minutes to hours depending on the severity and duration of the hypoglycemia. Death or permanent brain damage resembling stroke can occur rarely as a result of severe hypoglycemia. See hypoglycemia for more on effects, recovery, and risks.

Further therapy and prevention depends upon the specific cause.

Most hypoglycemia due to excessive insulin occurs in people who take insulin for type 1 diabetes. Management of this hypoglycemia is sugar or starch by mouth (or in severe cases, an injection of glucagon or intravenous dextrose). When the glucose has been restored, recovery is usually complete. Prevention of further episodes consists of maintaining balance between insulin, food, and exercise. Management of hypoglycemia due to treatment of type 2 diabetes is similar, and the dose of the oral hypoglycemic agent may need to be reduced. Reversal and prevention of hypoglycemia is a major aspect of the management of type 1 diabetes.

Hypoglycemia due to drug overdose or effect is supported with extra glucose until the drugs have been metabolized. The drug doses or combination often needs to be altered.

Hypoglycemia due to a tumor of the pancreas or elsewhere is usually curable by surgical removal. Most of these tumors are benign. Streptozotocin is a specific beta cell toxin and has been used to treat insulin-producing pancreatic carcinoma.

Hyperinsulinism due to diffuse overactivity of beta cells, such as in many of the forms of congenital hyperinsulinism, and more rarely in adults, can often be treated with diazoxide or a somatostatin analog called octreotide. Diazoxide is given by mouth, octreotide by injection or continuous subcutaneous pump infusion. When congenital hyperinsulinism is due to focal defects of the insulin-secretion mechanism, surgical removal of that part of the pancreas may cure the problem. In more severe cases of persistent congenital hyperinsulinism unresponsive to drugs, a near-total pancreatectomy may be needed to prevent continuing hypoglycemia. Even after pancreatectomy, continuous glucose may be needed in the form of gastric infusion of formula or dextrose.

High dose glucocorticoid is an older treatment used for presumptive transient hyperinsulinism but incurs side effects with prolonged use.

The following is a brief list of hormones and metabolites which may be measured in a critical sample. Not all tests are checked on every patient. A "basic version" would include insulin, cortisol, and electrolytes, with C-peptide and drug screen for adults and growth hormone in children. The value of additional specific tests depends on the most likely diagnoses for an individual patient, based on the circumstances described above. Many of these levels change within minutes, especially if glucose is given, and there is no value in measuring them after the hypoglycemia is reversed. Others, especially those lower in the list, remain abnormal even after hypoglycemia is reversed, and can be usefully measured even if a critical specimen is missed.

Part of the value of the critical sample may simply be the proof that the symptoms are indeed due to hypoglycemia. More often, measurement of certain hormones and metabolites at the time of hypoglycemia indicates which organs and body systems are responding appropriately and which are functioning abnormally. For example, when the blood glucose is low, hormones which raise the glucose should be rising and insulin secretion should be completely suppressed.

Normally, endogenous insulin production is suppressed in the setting of hypoglycemia. A 72-hour fast, usually supervised in a hospital setting, can be done to see if insulin levels fail to suppress, which is a strong indicator of the presence of an insulin-secreting tumor.

These blood tests are needed to diagnose insulinoma:

- glucose

- insulin

- C-peptide

If available, a proinsulin level might be useful, as well. Other blood tests may help rule out other conditions which can cause hypoglycemia.

Ketones in the urine or blood, as detected by urine strips or a blood ketone testing meter, may indicate the beginning of diabetic ketoacidosis (DKA), a dangerous and often quickly fatal condition caused by high glucose levels (hyperglycemia) and low insulin levels combined with certain other systemic stresses. DKA can be arrested if caught quickly.

Ketones are produced by the liver as part of fat metabolism and are normally not found in sufficient quantity to be measured in the urine or blood of non-diabetics or well-controlled diabetics. The body normally uses glucose as its fuel and is able to do so with sufficient insulin levels. When glucose is not available as an energy source because of untreated or poorly treated diabetes and some other unrelated medical conditions, it begins to use fat for energy instead. The result of the body turning to using fat instead of glucose for energy means ketone production that is measurable when testing either urine or blood for them.

Ketone problems that are more serious than the "trace or slight" range need immediate medical attention; they cannot be treated at home. Veterinary care for ketosis/ketoacidosis can involve intravenous (IV) fluids to counter dehydration, when necessary, to replace depleted electrolytes, intravenous or intramuscular short-acting insulin to lower blood glucose levels, and measured amounts of glucose or force feeding, to bring the metabolism back to using glucose instead of fat as its source of energy.

When testing urine for ketones, the sample needs to be as fresh as possible. Ketones evaporate quickly, so there is a chance of getting a false negative test result if testing older urine. The urine testing strip bottle has instructions and color charts to illustrate how the color on the strip will change given the level of ketones or glucose in the urine over 15 (ketones–Ketostix) or 30 (glucose–Ketodiastix) seconds. Reading the colors at those time intervals is important because the colors will continue to darken and a later reading will be an incorrect result. Timing with a clock or watch second hand instead of counting is more accurate.

At present, there is only one glucometer available for home use that tests blood for ketones using special strips for that purpose–Abbott's Precision Xtra. This meter is known as Precision, Optium, or Xceed outside of the US. The blood ketone test strips are very expensive; prices start at about US$50 for ten strips. It is most likely urine test strips–either ones that test only for ketones or ones that test for both glucose and ketones in urine would be used. The table above is a guide to when ketones may be present.

The use of an inexpensive glucometer and blood glucose testing at home can help avoid dangerous insulin overdoses and can provide a better picture of how well the condition is managed.

A 2003 study of canine diabetes caregivers who were new to testing blood glucose at home found 85% of them were able to both succeed at testing and to continue it on a long-term basis. Using only one blood glucose reading as the reason for an insulin dose increase is to be avoided; while the results may be higher than desired, further information, such as the lowest blood glucose reading or nadir, should be available to prevent possible hypoglycemia.

Urine strips are not recommended to be used as the sole factor for insulin adjustments as they are not accurate enough. Urine glucose testing strips have a negative result until the renal threshold of 10 mmol/L or 180 mg/dL is reached or exceeded for a period of time. The range of negative reading values is quite wide-covering normal or close to normal blood glucose values with no danger of hypoglycemia (euglycemia) to low blood glucose values (hypoglycemia) where treatment would be necessary. Because urine is normally retained in the bladder for a number of hours, the results of urine testing are not an accurate measurement of the levels of glucose in the bloodstream at the time of testing.

Glucometers made for humans are generally accurate using canine and feline blood except when reading lower ranges of blood glucose (<80 mg/dL), (<4.44 mmol/L). It is at this point where the size difference in human vs animal red blood cells can create inaccurate readings. Glucometers for humans were successfully used with pets long before animal-oriented meters were produced. A 2009 study directly compared readings from both types of glucometers to those of a chemistry analyzer. Neither glucometer's readings exactly matched those of the analyzer, but the differences of both were not clinically significant when compared to analyzer results. All glucometer readings need to be compared to same sample laboratory values to determine accuracy.

Children "outgrow" ketotic hypoglycemia, presumably because fasting tolerance improves as body mass increases. In most the episodes become milder and more infrequent by 4 to 5 years of age and rarely occur after age 9. Onset of hypoglycemia with ketosis after age 5 or persistence after age 7 should elicit referral and an intensive search for a more specific disease.

Serum glucose levels are measured to document the degree of hypoglycemia. Serum electrolytes calculate the anion gap to determine presence of metabolic acidosis; typically, patients with glycogen-storage disease type 0 (GSD-0) have an anion gap in the reference range and no acidosis. See the Anion Gap calculator.

Serum lipids (including triglyceride and total cholesterol) may be measured. In patients with glycogen-storage disease type 0, hyperlipidemia is absent or mild and proportional to the degree of fasting.

Urine (first voided specimen with dipstick test for ketones and reducing substances) may be analyzed. In patients with glycogen-storage disease type 0, urine ketones findings are positive, and urine-reducing substance findings are negative. However, urine-reducing substance findings are positive (fructosuria) in those with fructose 1-phosphate aldolase deficiency (fructose intolerance).

Serum lactate is in reference ranges in fasting patients with glycogen-storage disease type 0.

Liver function studies provide evidence of mild hepatocellular damage in patients with mild elevations of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels.Plasma amino-acid analysis shows plasma alanine levels as in reference ranges during a fast.

Hyperinsulinism may also refer to forms of hypoglycemia caused by excessive insulin secretion. In normal children and adults, insulin secretion should be minimal when blood glucose levels fall below 70 mg/dL (3.9 mM). There are many forms of hyperinsulinemic hypoglycemia caused by various types of insulin excess. Some of those that occur in infants and young children are termed congenital hyperinsulinism. In adults, severe hyperinsulinemic hypoglycemia is often due to an insulinoma, an insulin-secreting tumor of the pancreas.

Insulin levels above 3 μU/mL are inappropriate when the glucose level is below 50 mg/dL (2.8 mM), and may indicate hyperinsulinism as the cause of the hypoglycemia. The treatment of this form of hyperinsulinism depends on the cause and the severity of the hyperinsulinism, and may include surgical removal of the source of insulin, or a drug such as diazoxide or octreotide that reduces insulin secretion.

That spontaneous hyperinsulinism might be a cause of symptomatic hypoglycemia was first proposed by Seale Harris, MD, 1924, in "Journal of the American Medical Association".

Dr. Seale Harris first diagnosed hyperinsulinism in 1924 and also is credited with the recognition of spontaneous hypoglycemia.

Evaluation of a patient with suspected glycogen-storage disease type 0 requires monitored assessment of fasting adaptation in an inpatient setting.

Patients typically have hypoglycemia and ketosis, with lactate and alanine levels in the low or normal part of the reference range approximately 5–7 hours after fasting.

A glucagon tolerance test may be needed if the fast fails to elicit the expected rise in plasma glucose. Lactate and alanine levels are in the reference range.

By contrast, a glucagon challenge test after a meal causes hyperglycemia, with increased levels of plasma lactate and alanine.

Oral loading of glucose, galactose, or fructose results in a marked rise in blood lactate levels.

Without adequate metabolic treatment, patients with GSD I have died in infancy or childhood of overwhelming hypoglycemia and acidosis. Those who survived were stunted in physical growth and delayed in puberty because of chronically low insulin levels. Mental retardation from recurrent, severe hypoglycemia is considered preventable with appropriate treatment.

Hepatic complications have been serious in some patients. Adenomas of the liver can develop in the second decade or later, with a small chance of later malignant transformation to hepatoma or hepatic carcinomas (detectable by alpha-fetoprotein screening). Several children with advanced hepatic complications have improved after liver transplantation.

Additional problems reported in adolescents and adults with GSD I have included hyperuricemic gout, pancreatitis, and chronic renal failure. Despite hyperlipidemia, atherosclerotic complications are uncommon.

With diagnosis before serious harm occurs, prompt reversal of acidotic episodes, and appropriate long-term treatment, most children will be healthy. With exceptions and qualifications, adult health and life span may also be fairly good, although lack of effective treatment before the mid-1970s means information on long-term efficacy is limited.

Several different problems may lead to the diagnosis, usually by two years of age:

- seizures or other manifestations of severe fasting hypoglycemia

- hepatomegaly with abdominal protuberance

- hyperventilation and apparent respiratory distress due to metabolic acidosis

- episodes of vomiting due to metabolic acidosis, often precipitated by minor illness and accompanied by hypoglycemia

Once the diagnosis is suspected, the multiplicity of clinical and laboratory features usually makes a strong circumstantial case. If hepatomegaly, fasting hypoglycemia, and poor growth are accompanied by lactic acidosis, hyperuricemia, hypertriglyceridemia, and enlarged kidneys by ultrasound, gsd I is the most likely diagnosis. The differential diagnosis list includes glycogenoses types III and VI, fructose 1,6-bisphosphatase deficiency, and a few other conditions (page 5), but none are likely to produce all of the features of GSD I.

The next step is usually a carefully monitored fast. Hypoglycemia often occurs within six hours. A critical blood specimen obtained at the time of hypoglycemia typically reveals a mild metabolic acidosis, high free fatty acids and beta-hydroxybutyrate, very low insulin levels, and high levels of glucagon, cortisol, and growth hormone. Administration of intramuscular or intravenous glucagon (0.25 to 1 mg, depending on age) or epinephrine produces little rise of blood sugar.

The diagnosis is definitively confirmed by liver biopsy with electron microscopy and assay of glucose-6-phosphatase activity in the tissue and/or specific gene testing, available in recent years.

There are no known ways of preventing LADA type 1 diabetes, though some researchers believe it could be stopped at a very early stage if a diagnosis is made prior to the body's destruction of its beta cells.

Although many factors influence insulin secretion, the most important control is the amount of glucose moving from the blood into the beta cells of the pancreas. In healthy people, even small rises in blood glucose result in increased insulin secretion. As long as the pancreatic beta cells are able to sense the glucose level and produce insulin, the amount of insulin secreted is usually the amount required to maintain a fasting blood glucose between 70 and 100 mg/dL (3.9-5.6 mmol/L) and a non-fasting glucose level below 140 mg/dL (<7.8 mmol/L).

When liver cells and other cells that remove glucose from the blood become less sensitive (more resistant) to the insulin, the pancreas increases secretion and the level of insulin in the blood rises. This increased secretion can compensate for reduced sensitivity for many years, with maintenance of normal glucose levels. However, if insulin resistance worsens or insulin secretion ability declines, the glucose levels will begin to rise. Persistent elevation of glucose levels is termed diabetes mellitus.

Typical fasting insulin levels found in this type of hyperinsulinism are above 20 μU/mL. When resistance is severe, levels can exceed 100 μU/mL.

In addition to being a risk factor for type 2 diabetes, hyperinsulinism due to insulin resistance may increase blood pressure and contribute to hypertension by direct action on vascular endothelial cells (the cells lining blood vessels). Hyperinsulinism has also been implicated as a contributing factor in the excessive production of androgens in polycystic ovary syndrome.

The principal treatments of hyperinsulinism due to insulin resistance are measures that improve insulin sensitivity, such as weight loss, physical exercise, and drugs such as thiazolidinediones or metformin.

The blood glucose can usually be raised to normal within minutes with 15-20 grams of carbohydrate, although overtreatment should be avoided if at all possible. It can be taken as food or drink if the person is conscious and able to swallow. This amount of carbohydrate is contained in about 3-4 ounces (100-120 mL) of orange, apple, or grape juice, about 4-5 ounces (120-150 mL) of regular (non-diet) soda, about one slice of bread, about 4 crackers, or about 1 serving of most starchy foods. Starch is quickly digested to glucose, but adding fat or protein retards digestion. Composition of the treatment should be considered, as fruit juice is typically higher in fructose which takes the body longer to metabolize than simple dextrose alone. Following treatment, symptoms should begin to improve within 5 to 10 minutes, although full recovery may take 10–20 minutes. It should be noted that over treatment does not speed recovery, and will simply produce hyperglycemia afterwards, which ultimately will need to be corrected. On the other hand, since the excess of insulin over the amount required to normalize blood sugar may continue to reduce blood sugar levels after treatment has produced an initial normalization, continued monitoring is required to determine if further treatment is necessary.

Diabetes mellitus is characterized by recurrent or persistent high blood sugar, and is diagnosed by demonstrating any one of the following:

- Fasting plasma glucose level ≥ 7.0 mmol/l (126 mg/dl)

- Plasma glucose ≥ 11.1 mmol/l (200 mg/dl) two hours after a 75 g oral glucose load as in a glucose tolerance test

- Symptoms of high blood sugar and casual plasma glucose ≥ 11.1 mmol/l (200 mg/dl)

- Glycated hemoglobin (HbA) ≥ 48 mmol/mol (≥ 6.5 DCCT %).

A positive result, in the absence of unequivocal high blood sugar, should be confirmed by a repeat of any of the above methods on a different day. It is preferable to measure a fasting glucose level because of the ease of measurement and the considerable time commitment of formal glucose tolerance testing, which takes two hours to complete and offers no prognostic advantage over the fasting test. According to the current definition, two fasting glucose measurements above 126 mg/dl (7.0 mmol/l) is considered diagnostic for diabetes mellitus.

Per the World Health Organization people with fasting glucose levels from 6.1 to 6.9 mmol/l (110 to 125 mg/dl) are considered to have impaired fasting glucose. people with plasma glucose at or above 7.8 mmol/l (140 mg/dl), but not over 11.1 mmol/l (200 mg/dl), two hours after a 75 g oral glucose load are considered to have impaired glucose tolerance. Of these two prediabetic states, the latter in particular is a major risk factor for progression to full-blown diabetes mellitus, as well as cardiovascular disease. The American Diabetes Association since 2003 uses a slightly different range for impaired fasting glucose of 5.6 to 6.9 mmol/l (100 to 125 mg/dl).

Glycated hemoglobin is better than fasting glucose for determining risks of cardiovascular disease and death from any cause.