A1C chart: A1C to average blood sugar

Where the numbers in this chart come from

Every row is the same single equation, not a lookup table someone typed in: estimated average glucose in mg/dL equals 28.7 multiplied by the A1C percentage, minus 46.7. That relationship came out of the A1c-Derived Average Glucose study, which compared laboratory A1C against continuous and repeated fingerstick glucose measurement in 507 people and published the resulting regression in Diabetes Care in 2008. The mmol/L column is that same mg/dL figure divided by 18.0182, the molar factor for glucose, so the two glucose columns are consistent with each other rather than each coming from a separate equation.

What the chart cannot tell you

A regression describes a population, not a person. In the ADAG data, individuals who shared an A1C had mean glucose values that differed meaningfully from one another, so your own average may sit above or below the row you land on. Separately, the A1C measurement itself has a tolerance: NGSP certification allows a certified assay to differ from the reference method by roughly half a percentage point. Read together, those two facts mean this chart is good for orientation and poor for splitting hairs — a difference of a tenth between two results is well inside the noise.

Why some reports show mmol/mol instead

Laboratories in the United Kingdom, Australia and much of Europe report A1C in IFCC units, as millimoles per mole, rather than as an NGSP percentage. The two are related by the NGSP/IFCC master equation: IFCC mmol/mol equals the NGSP percentage minus 2.15, multiplied by 10.929. So a result reported as 48 mmol/mol is the same measurement as one reported as 6.5%. The third column of the chart gives that conversion for every row, which is usually what someone comparing an old report with a new one actually needs.

Reading a range rather than a line

The ADA defines three ranges on this scale: below 5.7%, the 5.7% to 6.4% band it calls prediabetes, and 6.5% and above, which is one of its diagnostic criteria for diabetes. Those are definitions of numeric ranges rather than conclusions about a person, and the ADA asks clinicians to confirm a result with a repeat test before diagnosing. That matters most exactly where people look hardest — at values sitting a tenth or two either side of a boundary, where the assay tolerance is wider than the gap.

Official sources

• Nathan et al., Translating the A1C Assay Into Estimated Average Glucose Values, Diabetes Care 2008
• American Diabetes Association — Standards of Care in Diabetes
• NGSP — IFCC standardisation and the NGSP/IFCC master equation
• NGSP — A1C method certification criteria

Frequently asked questions