Type 2 Diabetes Diagnosis
Type 2 Diabetes Diagnosis
Diagnosis in type 2 diabetes is made by a blood test and almost never by a symptom, which is what makes the thresholds worth understanding. This page sets out the three tests and what each one reflects, the values at which the diagnosis is made, why an abnormal result in someone with no symptoms is repeated before anything is concluded, the situations in which HbA1c cannot be used at all, and what the first assessment establishes beyond the diagnosis itself.
The three tests, and what each one reflects
Three tests can establish the diagnosis, and they are not versions of each other. Each one samples a different part of the day, which is why a person can be above the line on one and below it on another.
- HbA1c measures the proportion of haemoglobin carrying glucose. Red cells live for two to three months, so the figure reflects average exposure across that window. It needs no fasting and no preparation, which is why it is the test most often used first.
- Fasting plasma glucose is a single measurement taken after at least eight hours without food. It is simple and it is specific to one moment — the moment before breakfast, which is not where most of the day is spent.
- The 75 g oral glucose tolerance test measures glucose two hours after a measured glucose drink. It describes how the body handles a load, and it is the test used in pregnancy and the one that catches a pattern the other two miss.
A fourth measurement has a narrower role: a random plasma glucose, taken at any time of day, is enough to make the diagnosis in someone with the classic symptoms, because the symptoms are supplying the confirmation that a second test would otherwise provide. It is not used to screen people who feel well.
What none of these tests does is date the condition. A result above the threshold tells you where glucose has been sitting, not how long it has been there — which is why the first assessment looks for what the preceding years have already done, rather than assuming the clock starts at the diagnosis.
The thresholds, and the range below them
The thresholds below are the values the international guidance defines. They describe populations, which is the reason the next section exists.
| Test | Diabetes | Below the threshold |
|---|---|---|
| HbA1c | At or above 6.5% (48 mmol/mol) | A band below 6.5% and above the ordinary range marks raised risk; the cut-offs and the label for it differ between guidance bodies |
| Fasting plasma glucose | At or above 7.0 mmol/L (126 mg/dL) | A fasting value below 7.0 but above the ordinary range is called impaired fasting glucose |
| Two-hour value, 75 g tolerance test | At or above 11.1 mmol/L (200 mg/dL) | A two-hour value below 11.1 but above the ordinary range is called impaired glucose tolerance |
| Random plasma glucose | At or above 11.1 mmol/L (200 mg/dL) with classic symptoms | Not used for screening in someone who feels well |
Two cautions sit on that table. The first is that units differ between countries, and a figure quoted without its unit is not a figure: HbA1c is reported as a percentage in most places and as mmol/mol in others, and glucose as mmol/L or as mg/dL. The second is that a result in the band below the threshold is not a borderline case of diabetes; it is a finding in its own right, with its own follow-up, described on the monitoring page.
Where HbA1c cannot be trusted, the glucose tests are used instead rather than added to it. That is dealt with next.
Why a single result is repeated
An abnormal result in someone with no symptoms is repeated before the diagnosis is made. This is not caution for its own sake: the thresholds were set by weighing sensitivity and specificity across populations, and any individual result can fall on the wrong side of the line.
Three things move a result temporarily. Acute illness and infection raise glucose, and so does the stress response around surgery or a hospital admission. Corticosteroids raise it substantially and predictably. And there is ordinary measurement variation between laboratories and between samples. A single result taken during any of those describes the week rather than the person.
The usual response is a second test on another day — the same test repeated, or a different one from the three. Where two tests disagree, the one that is abnormal is the one that is repeated, and where disagreement persists the plan follows the pattern and the clinical picture rather than the higher figure automatically. In someone with the classic symptoms and a clearly raised reading, none of this applies, because the symptoms are doing the confirming.
There is also a group of situations in which HbA1c should not be used at all, and using it there produces a wrong answer in either direction. Anything that shortens red cell survival or changes haemoglobin lowers it: anaemia and its treatment, haemoglobin variants and thalassaemia, pregnancy, recent transfusion or blood loss, and advanced kidney disease. In those situations glucose measurements are used instead, and if the two disagree the glucose result is the one that decides.
Finally, a diagnosis should not be made for the first time during a severe illness unless it is unambiguous. Where the result was taken in the middle of an infection or an admission, it is retested once the person has recovered, and the interim is managed rather than labelled.
What the first assessment establishes
The diagnosis is the beginning of the first visit rather than the end of it. Three things have to come out of that visit beyond the label, and a consultation that produces a prescription without them has skipped the part that makes the prescription appropriate.
| Assessment | What it establishes |
|---|---|
| Dilated eye examination with retinal imaging | Whether retinopathy is already present. It is silent, it is common at diagnosis, and it is far more treatable when it is found this way |
| Urine albumin-to-creatinine ratio and estimated filtration rate | How the kidneys are now, and whether albumin is leaking — which changes the medicine choice as well as the follow-up |
| Foot examination — sensation, vibration, pulses | Whether sensation has already reduced and whether the circulation is adequate, both of which set the interval for checking from now on |
| Blood pressure, lipids, liver, weight and waist | Where the rest of the risk sits. Treating blood pressure and lipids usually reduces risk more than any further change in glucose |
| A review of every other medicine | Whether something else is raising glucose, and whether anything about to be prescribed will interact |
| Type established, with antibodies and C-peptide where the picture is not typical | Whether this really is type 2 — which decides the treatment rather than the label |
| Vaccination history, dental health, sleep and smoking | The parts of the plan that are not medicines and are usually the ones that get deferred |
| A target agreed and written down, with the screening interval | What the plan is aiming at, and when the eyes, kidneys and feet are looked at again |
What makes the visit useful is what arrives with it rather than what happens in the room: previous results with dates, the medicines actually being taken with their doses, blood pressure readings taken at home, and the reports of any eye or foot screening already done. Our endocrinology department carries the fundus screening device, the foot screening device and the lipid analyser used in this assessment, and the thyroid subspecialty is available where thyroid disease — which clusters with diabetes — needs looking at in the same visit.
Establishing that it really is type 2
Most people diagnosed as adults with type 2 diabetes genuinely have it. A minority do not, and the ones who are mislabelled are mislabelled in both directions, which is why the question is asked at the start rather than after a few years of the wrong treatment.
| Feature | Type 2 | Type 1 and other forms |
|---|---|---|
| Tempo of onset | Gradual, across months or years, often found on a routine test | Rapid, over days or weeks, usually because symptoms appeared |
| Body weight | Often carrying excess weight, particularly around the abdomen | Often lean, and weight loss before diagnosis is common |
| Ketones at presentation | Unusual, unless there is a severe illness alongside | Common, and ketoacidosis may be how it presents |
| Autoantibodies | Absent | Present in autoimmune diabetes — this is usually the deciding result |
| C-peptide | Normal or high at first, falling across the years | Low, because the body is making little insulin |
| Family pattern | Common in the family, but not across three generations in a clear line | A strong line across several generations suggests monogenic diabetes instead |
Three situations deserve to be looked for specifically. A lean adult diagnosed with type 2 who needs insulin much sooner than expected may have slowly progressing autoimmune diabetes, and antibodies settle it. A young, lean person with no insulin resistance and affected relatives in three generations may have monogenic diabetes, which changes the medicine and changes what is told to their relatives. And a person whose glucose rose with a corticosteroid course, or after pancreatic illness or surgery, has a secondary form, where the underlying cause is the thing to treat.
Where the picture turns out to be type 1 rather than type 2, the care is described on the type 1 diabetes page, which covers insulin, monitoring and the same complication screening from a different starting point.