Biology Reference
In-Depth Information
and especially its component HbA
1c
(Santiago [1993]). Since then, HbA
1c
has been confirmed as the gold-standard assay for people with T1DM
and T2DM. The guidelines specify that HbA
1c
of 7% corresponds to a
mean BG of 8.3 mmol/L (150 mg/dl), an HbA
1c
of 9% corresponds to a
mean BG of 11.7 mmol/L (210 mg/dl), and a 1% increase in HbA
1c
corresponds to an increase in mean BG of 1.7 mmol/L (30 mg/dl).
HbA
1c
, however, only captures average glycemia. The radical
fluctuations in BG levels represented in Figure 5-2—especially those that
take the BG levels into the hypoglycemic range—are very hazardous, but
are not recognized in the HbA
1c
test. In fact, the DCCT concluded in
1997 that only about 8% of future severe hypoglycemia episodes can be
predicted from known variables, including HbA
1c
. Predictions improved
to 18% with a more recent model using history of SH, hypoglycemia
awareness, and autonomic score (see DCCT Research Group [1997];
Gold et al. [1997]). Given that intensive therapy increases the risk for
hypoglycemia, strict control of T1DM implies that BG levels should be
closely monitored for large deviations at both the low and the high end
of the BG scale. It also follows that the risk for hypoglycemia needs to be
monitored by means other than HbA
1c
.
The rapid development of home BG monitoring devices provides a
means for monitoring BG fluctuations, and, in particular, monitoring for
hypoglycemia. Contemporary memory meters can store several hundred
BG readings and can calculate various statistics, including the mean of
these BG readings. Increasingly, research is focused on developing
devices for continuous, or nearly continuous, non-invasive
self-monitoring of BG. Two new journals, Diabetes Technology &
Therapeutics and Diabetes Science & Technology, were launched in 1999 and
2006, respectively, to report technological advances, including
information processing.
Still, surprisingly little attention has been devoted to processing or
mathematically interpreting these almost continuous data streams. This
deficiency is partially because of the only modest success achieved, until
recently, in predicting BG on the basis of previous readings. Although a
significant proportion of the variance of HbA
1c
can be accounted for by
BG readings, attempts to predict patients' vulnerability to SH were
particularly unsuccessful.
In contrast, a simple, recently developed mathematical marker, the low
BG index (LBGI), has predicted 40% of SH episodes in the subsequent
six months, using routine self-monitoring BG readings (Kovatchev et al.
1998). This improvement in predictive power originates from the use of
diabetes-oriented mathematical methods that take into account the
specific mathematical properties of the BG measurement scale. The use
of this diabetes-specific mathematical model can substantially improve
the forecasting of hypoglycemia and the overall quality of the
monitoring to control diabetes.
