A High Glucose on a Random Sample with a Normal HbA1c
A glucose taken at any time of day is a snapshot of what you last ate and when. HbA1c is the running average of the past three months. When the snapshot is high and the average is not, the average is the more informative of the two, and the usual explanation is simply that the sample was not fasting.
The pattern on your report
- Glucose High · mild Key
- HbA1c Normal Key
- Hemoglobin Normal Key
Printed as: Glucose in mmol/Lor mg/dL— An eighteenfold difference: 9.0 mmol/L is 162 mg/dL.Hemoglobin in g/Lor g/dLHbA1c in mmol/molor %— Not a simple multiplication. The conversion is (mmol/mol divided by 10.929) plus 2.15, so 48 mmol/mol is 6.5%. UK reports print mmol/mol, US reports print the percentage, and some print both.
Why the numbers look like this
Blood glucose is meant to move. It climbs after a meal, peaks somewhere around an hour later, and returns toward baseline over the next couple of hours, and none of that is abnormal. A random sample lands wherever it lands on that curve.
HbA1c smooths that out. Because glucose attaches to hemoglobin irreversibly and red cells live about three months, the proportion that carries a glucose molecule reflects the average exposure across that window. Individual peaks contribute, but one meal is diluted into ninety days of them.
Stress is the other reason a single glucose runs high. Illness, pain, injury and surgery all release cortisol and adrenaline, which push glucose up for as long as the stress lasts. That rise is real and it is not diabetes, though in some people it does reveal a reduced capacity that was already there.
Not being flagged is not the same as normal
There is no meaningful reference range for a random glucose, which is why the flag beside it carries so little weight. Diagnosis rests on defined tests: a fasting sample, a glucose tolerance test, or an HbA1c, each with its own threshold. US practice sets prediabetes at 5.7% to 6.4% and diabetes at 6.5%, while UK practice uses 42 to 47 mmol/mol for the high-risk band. A random glucose belongs to none of those schemes unless it is very high, at or above 11.1 mmol/L (200 mg/dL), and accompanied by symptoms, in which case it is diagnostic on its own.
What else on the report can hide this
The question that resolves this is what and when you last ate, and it is not on the panel. Ask it before ordering anything else.
Where the two genuinely disagree, look at the red cells before concluding anything about glucose. Anything that shortens red cell survival lowers HbA1c and makes it read falsely reassuring: hemolysis, bleeding, a transfusion, the later months of pregnancy, and advanced kidney or liver disease. Hemoglobin variants interfere with some assay methods outright. A low hemoglobin or an odd MCV alongside a mismatched pair is the signal to distrust the HbA1c rather than the glucose.
If steroids have been taken recently, they raise post-meal glucose disproportionately, which produces exactly this combination and settles after the course.
What usually causes it
Listed from most to least common — not from most to least serious.
- Very common
The sample was not fasting
By far the commonest explanation, and it needs no test to establish. A meal, a sugary drink, or even coffee with sugar in the hours beforehand. In hospital, a glucose-containing drip running at the time does the same.
- Very common
Stress or acute illness
Infection, pain, injury or surgery raise glucose for as long as the stress lasts. It resolves, though it sometimes reveals reduced reserve worth rechecking later.
- Common
Steroids
Raise post-meal glucose far more than fasting glucose, which is precisely this pattern. Recent courses count, not only current ones.
- Common
Early insulin resistance
Post-meal peaks appear years before the fasting number or the average moves. A glucose tolerance test is what shows it.
- Common
An HbA1c that reads falsely low
Hemolysis, recent blood loss, pregnancy, kidney or liver disease, or a recent transfusion. Here the glucose is telling the truth and the average is not.
- Uncommon
A hemoglobin variant interfering with the assay — in people of African, Mediterranean, Middle Eastern or Southeast Asian ancestry
HbS, HbC, HbD or HbE trait. The effect depends on the laboratory's method, so the laboratory is who to ask.
- Uncommon
Genuine new diabetes
Possible when the glucose is markedly high, when there are symptoms, or in type 1 developing quickly enough that the average has not caught up. Thirst, weight loss and passing large volumes of urine change the urgency entirely.
What is usually checked next
- A fasting glucose, or an HbA1c repeated in three months Puts the question onto a test with a defined threshold, which a random glucose does not have.
- Oral glucose tolerance test Catches post-meal rises that both the fasting number and the average miss, which is the hypothesis this pattern raises.
- Full blood count and ferritin Identifies the red cell problems that make an HbA1c read falsely low.
- A record of what was eaten and when, before the sample Answers the commonest explanation without a blood test at all.
When to seek care sooner
- Emergency Deep rapid breathing, vomiting, abdominal pain, or a sweet smell on the breath
- Emergency Drowsiness or confusion with a very high glucose
- Emergency Vomiting and abdominal pain on an SGLT2 inhibitor, even with a glucose that is not high
- Same day Heavy thirst, passing large volumes of urine, blurred vision, or unexplained weight loss
- Soon A random glucose that is very high on more than one occasion
Questions worth bringing to your appointment
- Was this sample fasting, and does that explain the result?
- Was I unwell or in pain when it was taken?
- Should I have a fasting glucose or a tolerance test rather than repeating this one?
- Could anything about my red cells be making the HbA1c read low?
- Have I taken steroids in the past few months?