Biology Reference
In-Depth Information
regulations that must be followed to ensure their protection and their
well-being.
Second, after we have the data and have computed the HBGI and LBGI,
what will the results tell us? How do we know whether our risk indices
provide a good fit to the clinical reality? This question is quite deep. It
encapsulates the essential difference between the risk index models
developed in this chapter and the models for tracking changes in
quantitative variables (such as population sizes and concentrations)
considered in Chapters 1 and 2. The ultimate test for a good fit was
made by estimating the deviation of the model values from the observed
data. Fundamentally, the difference stems from the fact the model
variables in Chapters 1 and 2 represented physical quantities measurable
(directly or indirectly) by means of standardized procedures. In this
chapter, however, we have stressed that the mean BG level over a four-
to six-week period cannot be estimated accurately from SMBG data, but
could be determined by a HbA
1c
test that is the standard for measuring
the average BG.
But how do we physically measure the risk for hypo- and
hyperglycemia? We have already emphasized that there are no
established standards. How can a model be validated in the absence of a
norm?
IX. VALIDATION OF THE BLOOD GLUCOSE
RISK INDICES
In the absence of a quantitative standard, new measures could be
initially validated by testing their ability to reflect verifiable medical
distinctions. For example, because of the physiological differences
between T1DM and T2DM, patients with T1DM are known to be at a
much higher risk of experiencing both hyper- and hypoglycemic
episodes, and their BG profiles are generally marked with frequent
excursions into the hyper- and hypoglycemic BG zones. In contrast,
the BG fluctuations of patients with T2DM have smaller amplitudes
and span a narrower range about the target zone, because large BG
fluctuations are caused by the instability of the insulin-glucose
dynamics. A critical factor for such instability is the insulin sensitivity of
the body, measured as the amount of glucose metabolized per unit of
insulin. Heuristically, it is natural to expect that insulin will produce a
stronger effect with higher insulin sensitivity (or lower insulin
resistance), causing larger BG fluctuations. Because T2DM is a disease of
increased insulin resistance, the BG fluctuations in T2DM are less
extreme (refer to Bergman et al. [1979] and Bergman [2003]).
Figure 5-10 graphically illustrates this difference. The data in
Figure 5-10(A) are from a 20-year-old man with T1DM from the age of 5.
The data in Figure 5-10(B) are from a 70-year-old man who had had
