Interpreting HbA1c values is tricky for several reasons: differing lab
measurements, variation among individuals, and misapprehension of the
relevant timeframe.

First trick: several different lab measurements have been introduced
since 1980, measuring slightly different subtypes with different limits
for normal values and thus different interpretive scales.

A National Glycohemoglobin Standardization Program began in 1996,
sponsored by the American Diabetes Association and others. See
reference 1. This program certifies HbA1c assays which conform to the
method used in the DCCT. However, as of 1998 other versions are still
in use in many places, both in the US and elsewhere. When you get a lab
result, be sure to look at what the lab considers to be the normal
range. Most discussion of HbA1c values in m.h.d appears to be based on
the DCCT, where the normal range is approximately 3.0-6.1%. Caveat
lector. (See part 5, Research, of this FAQ for more information on the
DCCT, the Diabetes Control and Complications Trial.)

Second trick: HbA1c levels appear to vary by up to 1.0% among
individuals with the same average bG. See reference 2.

This is very recent research and its implications are not yet clear. The
actual reaction rates governing the formation of HbA1c may vary among
individuals. Some of the variation may be due to differences in
erythrocyte (red blood cell) survival times -- the rough 90-120 day
range noted earlier -- although other work limits this to a small part
of the total variation (see reference 5). Variations in the HbA1c
formation rate may or may not correlate with the rate of damage to
other tissues.

While we await further research, we can only say that differences of
1.0% from one individual to another may not be meaningful.

Although HbA1c varies among individuals with the same average bG, it is
very stable for any given individual. Thus a change of 1.0% in your own
HbA1c is definitely meaningful.

Third and final trick: most medical professionals have been given
incorrect information about the timeframe which HbA1c represents.
Even textbooks normally state the 90-120 day average, as does the
American Diabetes Association in its Position Statement on Tests of
Glycemia in Diabetes (see reference 1).

The longer estimate is based on the assumption that the conversion of
hemoglobin A to HbA1c is essentially irreversible. This was a
reasonable assumption before the reaction rates were actually measured.
See the following section for information about the research which
measured the reaction rates and simulated the consequences.

See the following section for the references mentioned above.