Page 418 - Handbook of Biomechatronics

The Artificial Pancreas 413

constant. In general the scattering coefficient is dominated by particle size
and the refractive index of the sample so is not a strong function of the wave-
length or the absorption coefficient. This allows it to be treated as a constant
by the K-M model. Thus, to determine the concentration, the measured
reflectance is just transformed by the K-M model, and scaled appropriately.
An example of the reaction used by the Lifescan ONETOUCH and the
SureStep provides an indication of the complexity of the chemical process
involved in these strips. GOx catalyzes the oxidation of glucose to form
gluconic acid and H 2 O 2 . Blood oxygen concentration is much lower than
that of the air, so oxygen from the air must diffuse into the strip to bring
the reaction to completion. Peroxidase then catalyzes the reaction of the
H 2 O 2 with 3-methyl-2-benzothiazolinone hydrazone (MBTH) and
3-dimethyllaminobenzoic acid (DMAB) to form MBTH-DMAB that has
an absorption peak at 635nm. This, and similar reactions used by other
manufacturers requires between 3 and 15μL of blood, can be read using a
reflectance meter within 10–30s.
One of the confounding factors is the range of red blood cell concentra-
tions (% hematocrit) found in whole blood. To overcome this problem, the
ONETOUCH meter includes two LEDs operating at 635 and 700nm,
respectively. The absorbance peak at 635nm is fairly sharp, so the absor-
bance by the dye at 700nm is negligible. Therefore, the second reflectance
measurement at 700nm provides a correction factor in a band where oxy-
hemoglobin provides some absorbance. The two signals are multiplexed to
allow independent measurements using the same photometer.
All photometers measure the reflectance as a function of time to ensure
that the curve follows a prescribed path as the reaction progresses. Some cor-
rect for temperature that affects the reaction speed, as well.
Optical systems have slowly lost market share to biosensor
(electrochemical)-based technologies as the latter require less blood, are
faster and are easier to calibrate (Bronzino, 2006).

3.2 Biosensor-Based Glucose Monitoring

A biosensor is a device that uses a biologically derived sensitive element in
association with a physiochemical transducer to recognize an organic mol-
ecule. The process to achieve this involves firstly differentiating target mol-
ecules from other chemicals and then implementing the physiochemical
transducer and signal processor to convert the signal to a readable form.
The molecular differentiation element can include receptors, enzymes,