http://www.healthline.com/health-news/diabetes-google-develops-glucose-monitoring-contact-lens-012314

Yes I saw this. This inspired the post actually. I think this (just) fulfils the non-invasive requirement but it would not be popular among non contact lens wearers. Also does tear glucose concentration correlate with plasma concentration?

Yes. But the main question is about the accuracy of different methods.

http://www.cnoga.com/

http://www.integrity-app.com/the-glucotrack/

http://www.gluco-wise.com/

I should point out, Pitt Researchers originally came up with the contact lens glucose monitoring concept back in 2003. I was told about it during oh of my grad school interviews....I hope google bought the rights to the technology.

http://www.upmc.com/media/NewsReleases/2003/Pages/glucose-sensor-contact-lens.aspx

The salivary glucose concentration is higher in diabetic patients. The relationship between these salivary and plasma glucose concentration is positive in normal subjects vs. diabetic patients. However, the relationship between these two variables in each group of patients is low, does it mean that the correlation is not significant. Probably, the salivary glucose concentration can be used to to screen people for diabetes but not to detect levels of plasma glucose concentration.

The expression of glucose transporters in luminal membrane of ductal cells can be used to understand this relationship:

http://www.ncbi.nlm.nih.gov/pubmed/24499577

Scientifically this is feasible but there is a still gap for engineering solutions to realise this objective although many techniques claims their durable. I thought a key point to achieve/meet the objective of non-invasive monitoring of blood glucose is to  demonstrate the correlation between non-invasive measurement and clinical standards together with the sensitivity and selectivity.