Diabetes-related investigations - HbA1c
What it measures
- Non-enzymatic glycation of the N-terminal valine of the beta-globin chain
- Reflects mean glycaemia over the preceding ~8-12 weeks
- Weighted: ~50% from the last month, ~25% from months 2-3
- Reported in % (DCCT/NGSP) and mmol/mol (IFCC)
| HbA1c % | mmol/mol | Mean glucose |
|---|---|---|
| 6.0 | 42 | 7.0 |
| 6.5 | 48 | 7.8 |
| 7.0 | 53 | 8.6 |
| 8.0 | 64 | 10.2 |
| 9.0 | 75 | 11.8 |
| 10.0 | 86 | 13.4 |
- Rule of thumb: each 1% = ~1.6 mmol/L mean glucose
What it is not
- Not a measure of variability or hypoglycaemia - a "good" HbA1c can conceal swings between 2 and 20
- Not valid where red cell lifespan is altered
Epidemiology
- ~1.3 million Australians with diagnosed diabetes; ~500,000 undiagnosed
- Aboriginal and Torres Strait Islander people: ~3-4x prevalence, earlier onset
- Prediabetes (HbA1c 6.0-6.4%) in ~15-16% of adults
- Haemoglobinopathy prevalence matters for assay validity - high in South-East Asian, Mediterranean, African, Pacific and Middle Eastern communities
When it misleads
- Glycation rate is proportional to glucose concentration and to red cell exposure time
- -> Anything that changes red cell survival changes HbA1c independently of glycaemia
| Effect | Mechanism | Examples |
|---|---|---|
| Falsely LOW | Shortened RBC survival / inc turnover | Haemolysis, haemoglobinopathy, splenomegaly, blood loss, recent transfusion, pregnancy, erythropoietin, iron/B12 replacement, hydroxyurea, ribavirin, dapsone, advanced CKD |
| Falsely HIGH | Prolonged RBC survival | Iron deficiency, B12/folate deficiency, splenectomy, aplastic anaemia |
| Assay interference | Structural | HbS, HbC, HbE, HbF, carbamylated Hb (uraemia), acetylated Hb (high-dose aspirin), alcohol |
- Iron deficiency raises HbA1c and treating it lowers it - a fall after iron replacement is not improved control
- CKD pulls in both directions - shortened survival lowers it, carbamylation raises it
Diagnostic thresholds
Diagnostic thresholds
- HbA1c >=6.5% (48 mmol/mol) = diabetes
- Must be a laboratory (NGSP/IFCC-standardised) assay - point-of-care devices are not approved for diagnosis in Australia
- Confirm with a second abnormal test unless the patient is symptomatic with unequivocal hyperglycaemia
- 6.0-6.4% (42-47) = high risk / prediabetes
- MBS rebates one diagnostic HbA1c per 12 months in asymptomatic at-risk adults (AUSDRISK >=12)
When HbA1c cannot be used - use OGTT instead
When HbA1c cannot be used - use OGTT instead
- Haemoglobinopathy (sickle trait/disease, thalassaemia, HbE, persistent HbF)
- Haemolysis, recent transfusion or major blood loss
- Pregnancy - gestational diabetes is diagnosed by OGTT, never HbA1c
- Children and adolescents with suspected T1DM
- Suspected T1DM at any age, or acute/rapid-onset hyperglycaemia
- Symptoms of <2-3 months' duration may not have raised the HbA1c yet
- Advanced CKD/dialysis, HIV on antiretrovirals, recent iron/B12 therapy
Discordant HbA1c and glucose - work through
- Confirm with fructosamine or glycated albumin (2-3 week window, independent of red cell survival; invalid in hypoalbuminaemia, nephrotic syndrome, thyroid disease)
- Or CGM-derived glucose management indicator (GMI)
- Consider a haemoglobin variant -> HPLC / electrophoresis; a lab using a boronate affinity method is less affected
Targets - individualised, not universal
Targets - individualised, not universal
| Patient | HbA1c target |
|---|---|
| General T2DM | <=7.0% (53) |
| Short duration, long life expectancy, no CVD | 6.0-6.5% if achievable without hypoglycaemia |
| Recurrent severe hypoglycaemia, hypoglycaemia unawareness, advanced complications, frailty, limited life expectancy | <=8.0% (64), or symptom control only |
| On metformin alone | <=6.0% reasonable |
| Pregnancy (pre-existing diabetes) | <6.0-6.5% if safe |
Monitoring interval
- 3-monthly until stable, then 6-monthly
- No value measuring more often than 3-monthly - the analyte has not had time to move
When HbA1c is unreliable, monitor with
- CGM metrics - the modern standard where available
- Time in range 3.9-10.0 mmol/L: target >70%
- Time below range <3.9: <4%, and <3.0: <1%
- Each 10% inc time in range approximates a 0.5-0.8% fall in HbA1c
- Structured SMBG, fructosamine
- CGM is PBS-subsidised for all Australians with T1DM, and in selected other groups
Interpretation in practice
- A rising HbA1c with normal fingersticks -> ask about post-prandial excursions, adherence, and iron status
- A falling HbA1c with symptoms -> ask about hypoglycaemia, not success
Conditions that distort HbA1c
- Haemoglobinopathies - thalassaemia, HbS/C/E, hereditary persistence of HbF
- Iron deficiency, B12/folate deficiency, haemolytic anaemia
- CKD and dialysis; erythropoietin therapy
- Pregnancy - physiologically lower HbA1c from inc red cell turnover and haemodilution
- Splenectomy (raises) and hypersplenism (lowers)
- Ethnic variation - HbA1c runs ~0.2-0.4% higher in people of African ancestry at the same mean glucose
- HIV antiretrovirals (abacavir, protease inhibitors)
- Cystic fibrosis-related diabetes (*HbA1c insensitive - screen with annual OGTT*)
Benefit of lowering HbA1c
- Each 1% reduction in HbA1c -> ~37% dec microvascular complications (UKPDS)
- Legacy effect - early tight control confers benefit persisting 10+ years after the trial ends
- Macrovascular benefit of glucose lowering is modest and slow; ACCORD showed excess mortality with aggressive targets in established CVD -> the case for individualised targets
- HbA1c drift upward over years reflects progressive beta-cell failure in T2DM (~4%/yr loss of function)
- Variability matters independently - high glycaemic variability predicts hypoglycaemia and possibly complications at the same mean
What to monitor alongside
- Annual: eGFR + urine ACR, retinal screening, foot review, lipids, BP
- The HbA1c is the least important number on that list if the ACR is rising
🔒
13 more sections, plus exam facts
Premium unlocks every note across every specialty, and the full exam fact library behind it.
Get premium access