That is the first thing to understand about numbers in diabetes: every number is a piece of the picture, not the whole picture. The second is that numbers are useful only if you know which ones matter, what they measure, what the units mean, and what you are aiming for. This chapter is a practical guide to all of that. We start with the tests that diagnose diabetes, then look closely at HbA1c, the most widely used long-term measure, and finish with the targets that guidelines suggest, why they differ between organisations, and why a target is a starting point for a conversation with your care team rather than a grade.

4.1 The Tests That Diagnose

Four ways to find out

Chapter 2 introduced the thresholds. Here is how the tests themselves work, and what each does well or badly.

Fasting plasma glucose (FPG). A blood sample is taken after you have had nothing to eat or drink except water for at least eight hours, usually overnight.[@ada2026] It is simple and cheap, and it reflects the glucose the liver releases when no meal is arriving. Its weakness, as we saw in Chapter 2, is that it can miss people whose main problem shows up only after eating.

Oral glucose tolerance test (OGTT). After the same overnight fast, a first blood sample is taken, then you drink a solution containing 75 grams of glucose, and a second sample is taken two hours later.[@who2006] This is the test that detects impaired glucose tolerance. It takes about two hours, it is less convenient, and it is used less often in routine care, but it catches people the fasting test misses. It is also the standard way to screen for gestational diabetes, in a form your obstetric team will describe.[@ada2026]

HbA1c. A blood sample, with no fasting needed, measures the share of your haemoglobin that has glucose attached to it, which reflects average glucose over roughly the previous two to three months. It is convenient, but it can be misleading in some people, as we will see in section 4.2. The WHO accepts it as an additional test for diagnosing diabetes.[@who2011hba1c]

Random plasma glucose. A glucose measurement at any time of day. A value of 200 mg/dL (11.1 mmol/L) or higher together with classic symptoms of high glucose, such as increased thirst, frequent urination, and unexplained weight loss, is enough to diagnose diabetes.[@ada2026]

One abnormal result is not a diagnosis

Unless a person has clear symptoms with a very high glucose, guidelines call for confirmation: two abnormal results, either on the same blood sample (for example, an FPG and an HbA1c) or on two separate occasions.[@ada2026] The reason is practical. Glucose and even laboratory results vary from day to day, and a single abnormal reading should not, on its own, change your life. It should prompt a repeat test.

Two languages: mg/dL and mmol/L

Glucose is reported in one of two units, depending on where you live. Some countries use mg/dL (milligrams per decilitre), others use mmol/L (millimoles per litre). They are two ways of writing the same quantity, and you can convert between them by dividing mg/dL by 18 to get mmol/L, or multiplying mmol/L by 18 to get mg/dL. (The exact factor is 18.016, which is why the numbers in this book are converted with a calculator, not rounded by eye.) So 100 mg/dL is 5.6 mmol/L, and 126 mg/dL is 7.0 mmol/L.

HbA1c also has two languages. Older and most popular reporting uses a percentage (%), while the internationally standardised unit is mmol/mol. An international consensus statement described the standardisation and asked laboratories to report both.[@hanas2010] The two are related by a fixed formula, and a few examples are in the table in the next section. Whenever you compare numbers with a friend, another book, or a website, check the units first. A glucose of "6.8" and a glucose of "122" can be the same result.

Tests that tell you the type

The tests above tell you whether someone has diabetes, but not which type. As Chapter 3 noted, that sometimes requires extra tests: islet autoantibodies, which show autoimmune (type 1) diabetes and mark its early stages,[@insel2015] and C-peptide, which shows how much insulin the pancreas is still making.[@buzzetti2020] They are not needed in every case, but they are useful when the picture does not fit.

What about fasting insulin?

In Chapter 1, I suggested you might ask about fasting insulin. It is worth explaining what that can and cannot do. Fasting insulin is not part of the routine diagnosis of diabetes. One reason is technical: when a workgroup of the American Diabetes Association compared 12 commercial insulin tests, the results differed a lot from one method to another, with a median variation between assays of 24 percent and a range from 12 to 66 percent.[@marcovina2007] A result from one laboratory may not be directly comparable with another's. Researchers do use fasting insulin with fasting glucose to estimate insulin resistance in a formula called HOMA-IR.[@matthews1985] It is best seen as a research and specialist tool, not a number to chase.

Science Corner: How HOMA-IR is calculated The formula is: fasting insulin (in microunits per millilitre) multiplied by fasting glucose (in mmol/L), divided by 22.5.[@matthews1985] For example, an insulin of 10 with a glucose of 5.0 mmol/L gives 10 × 5.0 ÷ 22.5 = 2.2. Higher values mean more insulin was needed to hold glucose steady, which points toward insulin resistance. Because insulin assays differ between laboratories,[@marcovina2007] there is no universal cut-off that applies everywhere, and a single value should always be interpreted by a clinician who knows the laboratory and the person.

4.2 HbA1c: What It Shows and What It Hides

The long-exposure photograph

Imagine taking a photograph with a very long exposure. Anything that moves quickly disappears, and what is left is a blurred record of everything that was there over time. HbA1c works in a similar way. Glucose in the blood sticks to the haemoglobin inside red blood cells, and the more glucose there is, and the longer it is there, the more of the haemoglobin gets coated. Because a red blood cell lives for about three months, the proportion coated reflects the average glucose environment over that period.

{{fig:F4-A}}

Turning a percentage into a glucose number

Many people find it hard to picture what "7 percent" means. In a large international study, researchers worked out the relationship between HbA1c and average glucose measured over many weeks, and it can be expressed as a simple formula: estimated average glucose (mg/dL) = 28.7 × HbA1c − 46.7.[@nathan2008] The table shows a few values, together with the international mmol/mol equivalent for HbA1c.

Estimated average glucose values are those published in the American Diabetes Association's conversion calculator.[@adacalc] The mmol/mol values were calculated from the standard formula (%, minus 2.15, times 10.929).[@hanas2010] They are estimates: an individual's actual average can differ from the value predicted from HbA1c.

Notice something in the table: an HbA1c of 7 percent corresponds to an average glucose of about 154 mg/dL (8.6 mmol/L), which is well above what most people would call "normal" after a fast. HbA1c averages all of the day, including the peaks after meals.

When HbA1c can mislead

HbA1c depends on red blood cells, and anything that changes red blood cells can change the number without changing your glucose. A WHO consultation listed the main groups of factors:[@who2011hba1c]

• Red blood cell production. Iron and vitamin B12 deficiency can raise HbA1c, while treatments that boost red cell production, and conditions that increase the number of young red cells, can lower it.
• Haemoglobin variants. Inherited variants of haemoglobin (haemoglobinopathies) can push results up or down, depending on the variant and on the laboratory method.
• Red blood cell lifespan. Removal of the spleen can lengthen it and raise HbA1c. Conditions that shorten red cell life, including some inherited blood disorders and certain drugs, can lower it.
• Laboratory method. Some methods can be thrown off by other substances in the blood, and the WHO noted that some of these interferences are invisible in certain assays.

The practical message is this: if your HbA1c does not fit what your home readings or your symptoms suggest, or if you know you have anaemia, a blood disorder, or a haemoglobin variant, tell your doctor. Another test may be more reliable for you.

What HbA1c hides: the swings

The second limit is the one in the opening of this chapter. Two people with the same average can have completely different day-to-day experiences. HbA1c cannot tell whether your average comes from a steady line, or from wide swings that include dangerous lows. That is one reason continuous glucose monitors (CGMs), which we explore in Chapter 12, have become so valuable.

An international consensus on CGM data proposed a set of readings that go beyond the average. The most important is time in range: the percentage of time your glucose stays between 70 and 180 mg/dL (3.9 and 10.0 mmol/L). For most adults with type 1 or type 2 diabetes, the recommended goals are:[@battelino2019]

• More than 70 percent of time in range, which is more than about 16 hours 48 minutes a day.
• Less than 4 percent of time below 70 mg/dL (3.9 mmol/L), which is less than an hour a day, and less than 1 percent below 54 mg/dL (3.0 mmol/L), which is less than 15 minutes a day.
• Less than 25 percent of time above 180 mg/dL (10.0 mmol/L), and less than 5 percent above 250 mg/dL (13.9 mmol/L).

For older people and others at higher risk, the consensus suggests a more relaxed goal for time in range (more than 50 percent) but an even stricter one for lows (less than 1 percent below 70 mg/dL), reflecting the priority of avoiding hypoglycaemia over reaching a tight glucose target.[@battelino2019] The consensus also asks that at least 70 percent of readings be captured over a 14-day period for the report to be meaningful, and it describes a glucose management indicator (GMI), a way of estimating an HbA1c-like value from CGM data.[@battelino2019]

Science Corner: Why does time in range matter if I already have an HbA1c? HbA1c is an average, and averages hide extremes. A person who spends half the day very high and half the day very low can have the same HbA1c as someone who stays comfortably in the middle. The lows carry their own risks, which we discuss in Chapter 13. Time in range and time below range are ways of seeing both the average and the swings, which is why the consensus lists them together with HbA1c rather than in place of it.[@battelino2019]

{{fig:F4-B}}

4.3 Targets: Where Are You Aiming?

Different organisations, different numbers

A target is a number that is expected to balance two things: the benefit of lower glucose in preventing long-term complications, and the risk of the treatment itself, especially hypoglycaemia. Because organisations weigh these differently, their headline numbers are not identical. Here are three you may come across.

American Diabetes Association (ADA). For many non-pregnant adults, the ADA's targets are an HbA1c below 7 percent, before-meal (preprandial) capillary glucose of 80 to 130 mg/dL (4.4 to 7.2 mmol/L), and a peak after-meal glucose below 180 mg/dL (10.0 mmol/L) measured one to two hours after the start of the meal. The ADA also says that less stringent goals, such as an HbA1c below 8 percent, may suit people with certain risk factors.[@adaguide2023]

International Diabetes Federation (IDF). The IDF's 2025 recommendations for type 2 diabetes give a general HbA1c target of below 7.0 percent (53 mmol/mol), which should be personalised, balancing complications against hypoglycaemia and taking into account the person's characteristics. The IDF advises a higher target in older people and a lower one in the newly diagnosed. For people using self-monitoring, it suggests a fasting or pre-meal glucose of 4.0 to 8.0 mmol/L (70 to 144 mg/dL), and an after-meal glucose of 4.0 to 8.0 mmol/L as the preferred range, with 4.0 to 9.0 mmol/L (70 to 160 mg/dL) acceptable.[@idf2025rec]

American Association of Clinical Endocrinology (AACE). The AACE takes a tighter view for people who can achieve it safely: an HbA1c of 6.5 percent for most non-pregnant adults, with a fasting glucose that may need to be below 110 mg/dL (6.1 mmol/L) and a two-hour after-meal glucose below 140 mg/dL (7.8 mmol/L) to get there. Less stringent goals of 7 to 8 percent are recommended for people with, for example, a history of severe hypoglycaemia, hypoglycaemia unawareness, limited life expectancy, advanced kidney disease, or extensive other illnesses.[@aace2022]

The mmol/mol and mmol/L values for ADA and AACE were converted from the published mg/dL and % values. The ADA values are from the glycemic-targets guidance of its 2023 Standards of Care.[@adaguide2023]

Why the numbers differ, and what to do with that

The differences are small, and they are not a sign that one organisation is right and the others are wrong. They reflect a real trade-off. Lower glucose can lower the risk of complications, but chasing a tight target with medicines that can cause lows carries risk of its own. That is why every one of these bodies says the target must be individualised: age, how long you have had diabetes, other illnesses, how easily you notice a low, and your own preferences all matter. Someone newly diagnosed with type 2 diabetes and no complications, managing with food and movement alone, may reasonably aim tighter. An older person with several health conditions, or someone who has had severe lows, is often better served by a more relaxed goal.

Science Corner: A target is not a grade If your HbA1c is 7.4 percent and your target is below 7 percent, that does not mean you have failed. It means there is room to review what is working with your care team, for example your food, your movement, or your medicines. If your HbA1c is 6.2 percent and you are having frequent lows, that is also worth reviewing, because a lower number is not automatically a better one. The direction and the trend, and the safety of how you got there, all matter.

The other numbers on your report

Glucose is only one set of numbers that matter. Blood pressure, cholesterol, and kidney function each have their own targets and are part of the same protective picture. Chapter 14 covers these, and the lab reports and log sheets in Part 6 give you a place to record them. For now, the important point is that a well-managed person is not defined by a single number, but by a set of them tracked over time.

Key Takeaways

• Four tests can diagnose diabetes: fasting glucose, the two-hour oral glucose tolerance test, HbA1c, and a random glucose with symptoms. A single abnormal result usually needs confirmation.
• Always check the units. Glucose is reported in mg/dL or mmol/L (divide mg/dL by 18 to convert), and HbA1c in % or mmol/mol.
• HbA1c reflects average glucose over roughly two to three months. It can be unreliable in iron or B12 deficiency, haemoglobin variants, and conditions that change red blood cell life.
• HbA1c cannot show swings. Time in range (more than 70 percent between 70 and 180 mg/dL for most adults) and time below range give a fuller picture.
• ADA, IDF, and AACE targets differ slightly (below 7 percent for ADA and IDF, 6.5 percent for AACE if it can be achieved safely), and all agree that targets should be individualised.
• Fasting insulin and HOMA-IR are specialist tools, not routine numbers to chase.

Action Points

1. Know your last three numbers and their units. Find your most recent fasting glucose, HbA1c, and (if you have one) time in range. Write down the date and whether they are in mg/dL or mmol/L and % or mmol/mol.
2. Ask for your estimated average glucose. If you have an HbA1c result, ask your doctor to explain it as an average glucose, and compare it with what you see at home.
3. Ask whether your HbA1c can be trusted. If you have anaemia, a known blood disorder, or a haemoglobin variant, ask whether another test should be used alongside it.
4. Agree on a personal target with your care team. Write down the number, the reason for it, and when you will review it.
5. If you use a continuous glucose monitor, ask for your time-in-range report, and note the percentage in range, below range, and above range.

This book is intended for education and does not replace personal medical advice. If you have diabetes or take glucose-lowering medication, please consult your healthcare team before changing your diet, exercise, or treatment.