Near-infrared, as you have no doubt heard, is supposedly a technology that allows the taking of a blood glucose reading without the need for a blood sample. The theory is simple: when you shine light on your hand, it passes through it and comes out on the other side (take a flashlight and put it up to your hand: if you look at it from the other side you’ll see your hand glowing red). By analyzing the changes in the wavelengths of light after it has passed through an object, it is possible to determine the chemical makeup of that object. Every substance has a spectrum of light that it will absorb and emit; so in theory you can tell just how much glucose is in the blood by analyzing the intensity of the spectrum it emits.
It’s not quite that easy, however (after all, if it were we’d already be using non-invasive meters). Glucose makes up a very small part of the bloodstream, so you have to screen out the much more intense spectrum of blood plasma, and red blood cells, and white blood cells, and so on, to say nothing of the spectra of fat, muscle, bone, skin, etc. Combined with the fact that many of these spectra overlap in places, you can see that obtaining an accurate reading is incredibly difficult.
Chemometrics, an algorithm that uses a combination of statistics and spectral analysis, can be used to sift through the spectra and determine which one corresponds to blood glucose. An important feature of using near-infrared light to make the reading is that the spectrum emitted by glucose is at its strongest compared to the other spectra, which makes analysis by chemometrics that much easier.
Easier, though, is a relative term. Even under the optimum feasible conditions created by using near-infrared light, the glucose spectrum is very weak and obtaining a reading that is accurate enough to give an exact (or even a close approximation) of a person’s glucose level is incredibly difficult. There are other problems, too. Research has indicated the infrared monitor’s calibration, which requires taking both near-infrared and conventional blood readings of a person’s glucose level over a period of weeks and under various controlled conditions, may need to be changed for different groups of people, or may even require recalibration for each individual. As you can see, there are a number of barriers between us and a non-invasive meter using near-infrared technology.
Even so, the prospects are far from bleak. Most experts not only believe that near-infrared sensors will work, but their arrival on the market is simply a matter of time. The question they disagree upon is how much.