The paper is way overhyped. They've just built a meta material near field antenna. Nothing new I can see.

There are other papers available via Googling that glucose can be measured at around 4.2GHz. Seems hard to do it reliably, though.