I have been reading Heinlein and Bradbury since I was a teenager.
never was a Trekkie but always paid attention to the tech.
in the future one only has to use a gizmo that looks like a cross between an
electronic stud finder and a OBD probe, beeps and tells you all you need to know.
that's what we need.
no needles. Just have to wonder how to test the blood unless you
can get into the blood.
or do it some other way.
a breathalyzer that tests for sugar??
how about a mouth swab or spit?
something that doesn't poke holes in us...
I think the short answer is that blood is really the only viable substance, for a few reasons. To make matters worse, the best blood sample is from an arterial sample, not a vein, and not a finger stick.
The difficulty that I have seen with other samples is that they don't reliably work, and often lag blood glucose by hours in most cases.
Some dogs can detect high and low blood sugar in
some people, but it is a bit of an open question whether they do it by human behavior or smell. It turns out that a number of conditions besides high blood sugar can produce volatile chemicals like ketones. Plus things like individual idiosyncrasies like the microflora in a person's nose and mouth can alter odors, ditto skin.
Personally, I think the through skin optical methods are not likely to work, ever, since there are so many proteins that have glucose tacked on (glycosylated, e.g. HbA1c). One ends up with a spectral signal that looks like glucose, but it is not the blood glucose, it is the labeled proteins. I've seen smart folks give it a good try.
The microneedles or electro-osmotic methods, at least so far, seem to suffer from something like a diffusion effect, and can differ from blood values as the blood values seem to get smoothed out, missing highs, and lows, and lagging by long periods of time as the glucose diffuses into the outer layers of skin. So, glucose spikes and lows do not show up well.
Even the sampling site makes a difference;
That's an approval submission plot, and please note the permitted errors of +/-15% below 75mg/dL, and +/-20% over 75mg/dL, all of the values are just comparing stick locations.
Here's a CGM error plot (CGM vs arterial sample);
(From
https://www.liebertpub.com/doi/10.1089/dia.2024.0035)
I think the best bet for the foreseeable future is a CGM, but I do remind myself that even those aren't perfect. They are way better than not testing frequently in my opinion.
Sorry. I could of course be wrong, and I would love to see a non-invasive technology work, but personally, I am not backing that with money, and I would not bet my health and life on it.
All the best,
Peter
More here for the interested
www.biolinq.com
