Michelle Bonn I believe Google has the right mix of financing, dedicated R&D people and high visibility in the market to launch the new lens. Obviously, regulatory safety, design and manufacturing processes need to be finalized before the product can ever be launched....but most product companies go through this stage. Google is in a unique position to leverage their branding to obtain the right money and resources to bring such a disruptive device to market. If only we all had the same pathway ahead of us.

Even if this exact device doesn't reach the market, some new form of glucose monitoring probably will.

Jerry Robinson Why assume immediately for the US?

So... which country, which market??

Paul M. Stein Michelle, I totally agree. I think technologically it would just be a matter of time before it would work well. And, it might be a great product for the Type I diabetic market. For Type II's, however, I wonder if the cost and compliance would be worth it for them. The question at hand is, what do they all do now for monitoring? Pop that pill every day and only occasionally test? If that's the way it is, then the market may not be huge. But, since Google was never in medical devices before, whatever the final market size is might just be fine.

thomas blank Glucose in tears is really lagged vs the blood, and the concentration is 5x to 10x times lower in tears than the blood glucose(individual dependence expected). I have read some amazing enzyme electrodes that use nanotechnology which are much more sensitive and may work at low glucose concentrations, but the lag and individual factor correction still exist for tears and its hard to imagine closing the loop for type one diabetics or even using it manually to direct an insulin pump. If another tech of 5x larger size manages to be notably more accurate it will kill this technology for diabetics as accuracy is more important to a type I diabetic than size. I think Google will skirt the FDA regs and put something out there as a consumer electronics product with applications including glucose. But I think that diabetics will quickly figure out the accuracy limitations and for them it may seriously limit the usefulness of the technology. But "wearable medical devices" don't need a successful application for diabetics to be profitable. I think fitness is going to make plenty of money with a much wider customer base, and those applications do not need much in the way of FDA regulatory approval. I more view these wearable devices as gadgets, not on the same level as devices developed under regulatory approval. But in 5-8 years and a few Billion dollars later who knows. This kind of money has never been spent before to my knowledge on new medical devices. And surely most of the likely tech that will be initially used in fitness is not new, but established tech that will have a minimal to moderate performance degradation when miniaturized and designed to be aesthetically acceptable to the wearable crowd. It will be interesting for sure as Google and Apple have plenty of money to fund these wearables...

No I do not believe it will succeed. It is not a matter of money - it has been tried for years by the some of the best med engineers/scientists with no FDA approval achieved.

Glucose in sweat has a better correlation to blood than glucose in tears, and has a shorter time constant which is critical for Type 1 diabetics. Best to expect would be a glucose indicator - not an FDA approved glucose reading for insulin therapy.
RNB.

Yes -- there is good data to show that Glucose concentrations in Tears do not track the plasma concentrations of large vein blood - there is considerable lag and peak concentrations in short term peaks are not reflected. Even capillary blood drawn say near the elbow or abdomen does not reflect the plasma concentrations well -- this is particularly so when one needs 2 hr post-prandial values which relicts the availability of insulin in the body. This is also the case with saliva which was once touted in the same lines, if I remember correctly.

So the issue is more biological / medical -- can we produce correlation data and algorithms for populations -- between glucose concentrations in tears with that of central venous blood plasma -- which will enable good and effective clinical management of diabetes. Maybe the availability of such a device may lead to the generation of data / correlations -- but this could take much more than a couple of years to happen and are investors / industries willing to invest time and money ??

Google has the funding and the dedicated talent to do this if anyone can--but I am aligned with Mr. Blank on this one. In the early to mid-2000s several different university investigators looked at using tears to monitor glucose levels. My understanding at that time was that the low concentration, the time lag compared to capillary blood values, and the significant variances in levels between individuals caused each investigator to eventually drop their quest--and since then FDA has further tightened the precision and accuracy requirements to be approved. So this will not be easy and take a very large cohort of test subjects to develop the calibration routines.

I look at it differently, when you go by what people said earlier of use of Telephone, Computers etc. Today most of the Medical Equipment are driven by new hardware and software - a domain Google is strong in.
So if they have the correct people in place for this, then it would work. Technology is helping in locating sensitive parameters with ease.

Donna Cusano Excellent comments all! Thomas, Don and Gobi point out the equivalence problem, but if it can be overcome, Google certainly has the funding to do it. Or the contact lens path will lead elsewhere. If you would like to cross-post to my article as a comment, please feel free to do so. http://telecareaware.com/google-contacts-for-diabetics-in-research/

Mark Iverson I too am quite familiar with all the efforts with noninvasive glucose measurements. As several others have pointed out, the time lag is very likely too long to be of much use. If it was your life depending on it, would you buy it? We're using an RF-based method which resulted in excellent prospective accuracy over 4 weeks with NO recalibrations. 5 diabetics and over 850 samples provide 'a compelling case' as one optical spectroscopy expert put it! It's about time to try a different approach... details and a link to website with results on my profile page. B Well!

Accuracy is key for the device to successfully navigate the regulatory pathway and gain consumer & physician acceptance. As stated earlier, there are many challenges with measuring glucose in tears ... in addition to the time lag and inter-individual variability, there also is the matter of intra-individual changes in tear layer composition in response to various stimuli. None of it will be easy, but Google is to be commended for their drive and willingness to journey into this challenging area.

As a diabetic who happens to be a designer, I get excited every time I hear about a contact lens that can tell you if your blood sugar level is too high or too low (this has made the news before). But this time, I have to admit that my first thought was that with Google's name applied to it, I don't want to be anywhere near it. I have a Nest, and wasn't too terribly upset when Google bought them, despite the privacy advocates telling me that now Google will know when I'm home. That doesn't bother me. But my medical information? That's massively different. Even if this does make it to market, the question is, will the market accept it?

Carol David Daniel I am with Robert Blair on this one, you only have to read this reference to see the litany of failed attempts to measure blood glucose indirectly: http://www.mendosa.com/noninvasive_glucose.pdf . We have our own project in this space to measure the glucose concentration directly and non-invasively using an optical fibre spectroscope coupled to a patented technique for removing the noise from light scatter in tissue and blood.

Willi Glettig Interesting, - Money is obviously a big issue in this diagnostic field and Google most probably has the money but so their major competitors e.g. Roche and others.

We have seen many non-invasive glucose measuring technology but all failed due to poor stability and reproducibility. The current invasive technology is reliable, unpleasant but affordable. New technology would have to compete with this status quo.

Although I am not an expert in this field I consider the contact lens strategy as too complicated for Google. Assuming that the contact lens is only the carrier for say glucoxidase chemistry (which by the way suffers stability problems) then the question arises, - what happens to patients that wear currently individually corrected contact lenses? Such a system would also need electronics to evaluate the electro-chemical data and to react in some form to it. Would such electronic be paced into the lens or connected in some form to the contact lens? I believe there are too many technical challenges. These will take many years to be sorted out.
I don’t think IT companies do have the tenacity to invest for say another 10 to 20 years to get a product for a potentially large but very competitive market. I suspect Google will start the project with an aim to sell the venture or IPR in say 2 to 3 years to a large diagnostic specialist.

Mark Iverson I have been involved with noninvasive glucose on and off for 10 years, starting in the early 90s.
Optical methods for noninvasive glucose have consumed over 95% of the startup $ for the last 30 years... and some of the brightest scientists and engineers. ALL of them have failed. The latest, C8Medisensor, failed after devouring over $120M, and even after it had obtained CE Mark. Optical tech was great for pulse oximetry, but it will never achieve the technical requirements (sensitivity, repeatability, specificity) for glucose. It is time to look at other technologies, and there are numerous advantages, both biz and tech, to our use of RF.

John Bennett MD Has Killer App written all over it; a non-invasive measuring of serum glucose will have diabetics the world over celebrating.

Maybe not with tears, for the above-cited reasons. My bet is that optics will find the solution. Don't know how will do it, just look what else has been discovered lately with optics
Hemoglobin can now be measured with
an iPhone and its camera (Hb Helper, not FDA approved: pending, in iTunes now; http://www.youtube.com/watch?v=NuHtY_l0EFQ&feature=share), as well as blood oxygen
by an app, iPhone, and iPhone camera (also, in iTunes store, but FDA approval pending; called Digi Doc at http://www.digidoctech.no/). I interviewed both those guys, the first from India, and the second from Norway, and they used similar technologies, both basically optics. Tried to understand that difficult science but could not even understand "Optics for Dummies", so I will leave it to others to understand!

Frans Luijendijk It sounds promising, but it should 'look' promising. No data seen on accuracy and time lag.

Lou Cabana Having worked in field of self monitoring blood glucose devices for a number of years, the need for a non-invasive device remains paramount. ( ask any parent who has to wake his child for a finger stick !) However as others have noted the technical challenges remain high ( the Mendosa link is a must read). Optical methods generally lack the specificity required as other soluble compounds will cause interferences. Even measuring glucose directly in blood is impacted by the presence of such common chemicals as aspirin, other sugars, uric acid to name just a few.of Given the new guidelines for accuracy and precision soon to be in effect ( ISO -Europe ISO15197:2013, as well as soon to be released FDA guidelines will be a high hurdle. Most of the current devices on the market will not meet these standards, so to expect a new technology to not only be as good but better is a huge hurdle to overcome. That being said, I would not bet against Google or the Microsoft related effort currently underway. I personally am excited about having Google in this space

It's about time to try a different approach. We are interested in different areas to avoid the existing principle of invasive sample taking.

David Austin PhD I agree Achim, invasive technologies are often counter productive for a diabetic monitoring their glucose levels.
However, having many years experience and expertise as a clinician and scientist in contact lenses and materials this is a very indirect approach along with many issues relating to mechanical, physiological and infection within this at risk group.

I'd love to see JNJ help partner on developing this concept given that they have deep experience with both glucose monitoring (Lifescan franchise) and contact lenses (Vistakon franchise). Could be a great convergence. Accuracy requirements might not need to be as great as other systems given fact that as a continuous monitor it would be more for focusing on trends rather than absolute values. If lens detected a potentially troubling trend you could always revert to finger stick on limited confirmatory basis.

Glenn Neuman Not likely to meet FDA's accuracy requirement. Aside from the plasma-to-tears time lag, I see the lens having more value as an acute rather than a trending device, since removal of the lens could create large time gaps. And why add the risk and regulatory complexity of a contact lens (PMA) to a glucose measuring device (510(k))? Mobile medical aps are inevitable and will greatly benefit the public health (and burden the FDA) but I'm not thrilled to see Google involved -- the lens will soon be cluttered with ads!

Frans Luijendijk Accuracy of a continues device is as important as for a regular monitor, especially in the lows. Yes trends are absolutely very very helpful but you also want to counteract in time on a BG that drops fast. Continuous monitoring is changing the way diabetes will be managed. Accurate information is crucial then.

thomas blank Lou, I am going to respectfully disagree on the interferences in the use of optical methods as being the limitation. One can measure benzene concentrations as low as 0.2% in gasoline -a sample with hundreds of interferences- using high resolution vibrational spectroscopy. the problem with skin is that it is heterogeneous and the optical volume must be reproducible. If you cant reproduce your sample volume no method will work. And consider that even pressure changes the optical volume as light scattering changes with strain on the collagen. Also ISF is pushed around with contact as it exists at low pressure in skin(<10torr). Almost every research attempt failed to consider or control these issues. Many attempts at optical methods were driven by people who knew how to design and build instruments but knew very little about skin or making complex analytical measurements, and this continues today. Also in my opinion many small startups have mediocre technical management, most don't understand the nature of research and pressure from investors requires them to claim to be ahead of where they truly are. Once this happens, it kills off methodical research necessary to solve difficult problems.

As for ISO 15197 2013, that will be applied to point sensors and yes it will probably knock quite a few finger stick meters off the market. If they apply it to CGM's, ALL of those will have to be removed from the market. The easiest path to an FDA approved noninvasive glucose monitor is via the continuous route. Currently only the Raman device of C8 could be a feasible continuous optical technology capable of resolving glucose as it is a very small size. And apple has hired some of those C8 folks so who knows... I worked for C8 as a senior data analysis engineer, so I know something of the data quality, and I think you might be surprised. The failure at C8 was a management and experience failure, not a technical failure. Management failed to properly shepherd the technology to a state where it could be shipped, none of top technical management had any medical device experience, most were from the communications industry. This will be the challenge of Google and Apple, can they properly manage the development so something this complex outside their core technical expertise.

Glenn Neuman And when I wore contacts -- I always had eye drops handy. Eyes get dry, irritated, get a painful spec of something caught under the lens -- the first response is to flush with drops. What will that do to the glucose concentration? Sorry guys, thumbs down from me on this one, but keep at it -- the future of non-invasive diagnostics is bright!

Bruce Flora Having worked on blood glucose meters for several years, I'm sure this is technically feasible.
Saliva and interstitial fluid had measurable glucose levels. (ISF was sensitive to timing, so the comment about timing is significant.)
What I'm not sure of is the marketing. Current practice is to give away the meter and recover the cost in disposable strip sales. I don't see a disposable in this system so a new kind of business model is needed. This should be interesting!

Walt Bassett This is old news, Several Contact lens companies have developed this technology over YEARS AGO..... and never went to market
Consider the business plan. The average person does not want to put a lens in their eye.
Most people who need eye correction won't even use contacts. Getting stuck with a pin is more palatable than sticking a dirty finger in your eye. This concept died on the vine after years of heavy investment due to much simple existing
...former Syntex Opthalmics

John Bennett MD Thanks Lou for the feedback about Optics; it is a tough field to understand, tried Stan Giblisco's book, a variation of the Dummy Brand, but no go.
Yes, having Google in this space is good, but no guarantee, as their foray into Google Health shows, as does their format of Google Hangouts, with the tech of making a videoconference generally not being used by a mass audience, although it does have a future. Hope they keep working on that one, and that they hone in on non-invasive monitoring of glucose.

Personally I believe the concept can work. Google is not the only player and competition will be tough. The technology is getting close to where it needs to be. My guess is a disposable lens, to be replaced every week, calibration during night while cleaning and aligned to blood measurement. Timelag is there, I believe with intelligent algorithms you can still extract valuable information for guidance.

I seriously doubt the value due to the technical reasons already listed-especially the time lag.

Since the most successful prior-art non-invasive test is urine glucose, there needs to be data showing that glucose in tears tracks blood glucose levels significantly better than urine glucose. Are there any studies showing this?

Patrick Mize, PhD Why does this have to be an electrode and tears just because this happens to be the first design that they are talking about? You have a window into the eye [the pupil] which you are putting a lense over. Any other optical / fluorescent technique could be used to look into the eye and measure glucose and other interesting biomarkers....

Anthony Antonuccio I think that Carolynn and Willi summed up the answer to the question Donna asked. I agree with those that answered 'no' to this question. Google's core business has and likely always will be in big data gathering and tracking analytics, for their advertising revenue is dependent on it. Yes, they have the funds to dabble in many areas and may influence. They have also spent lots to follow others as a distraction. Are they likely to be a market leader in smart contact lenses for diabetics, and will their product see the light of day in the market? I highly doubt it, and I'm an optimist by nature.

David Austin PhD I agree with Patrick, there are more obvious methods to monitor glucose levels, already this has been good publicity for Google if you think of all the links made by different media sites.

Mark Iverson Thomas Blank stated, "the problem with skin is that it is heterogeneous and the optical volume must be reproducible. If you can't reproduce your sample volume no method will work."
Thomas is correct when talking about optical methods since the beam is small and 'interrogates' perhaps a cubic mm of tissue. This is why optical methods are so very sensitive to positioning and pressure and other external factors before you even get to the living tissue. Our tech is NOT optical, and interrogates several cubic *centimeters* of tissue, and thus we get much better repeatability. Optical methods also have very poor SNR, and therefore long signal integration times, so they have been limited to 'continuous' sensing where they are attached to the body. Our scans take less than 5 seconds and are capable of both spot and continuous sensing. These are just two of the technical advantages that RF has over optical methods for this kind of measurement.

When we first started trying to model the in-vivo data, we could get good accuracy during a day, but not several days, let alone weeks; a common problem regardless of technology. Once we determined what was causing the day to day difference, our predictive accuracy became weeks; even out to 4 months with only slight degradation. ANY noninvasive tech that can't show predictive (prospective) accuracy over at least a few weeks, with NO recalibrations and NO cherry-picking of data, is still in the research phase. Our data is available for anyone to review; I'd like to see how well Google's technology does with predictive accuracy over time... that's the real test!

David Lim, Ph.D. RAC, CQA There are user needs based on intended use. As for developing a medical device, it can be screening, monitoring and/or diagnostic purposes. Technical challenges/requirements differ depending up the intended use. Under the circumstances when all things considered, it warrants that Google should invest time and money to show/nail down technical parameters/limitation (at least up to feasibility studies) in comparison with those using blood plasma (e.g., the same or equivalent level of clinical sensitivity and clinical specificity -achievable or if not, at what extent?). Then decide for the next step!

As others have pointed out, this is a great case of big technical challenges vs massive available R&D budget. To me, it looks a lot like Google has put an interesting tail on a very big kite and sent it soaring into the media skies to see who notices. Maybe if they DO bring this to market, the lenses can be delivered by Amazon's drone fleet.

thomas blank Mark, the sample size depends on the optical method and the optical sensor. NIR can interrogate a path of cylinder with height of 1-1.5mm and diameter of 4mm, far more than a cubic mm. Yes those who have tried with a 1mm3 volume probably really struggled. Human tissue is most heterogeneous in depth as it is layered and there is also lots of back scatter at layers. The RF method sounds interesting, though there is a question whether it is specific to glucose or whether water shifts or some other analytes impact the signal. A diabetic on diuretics, as many late stage insulin dependent diabetics are, will experience large water shifts not related to glucose. And yes, I agree a single calibration must work for years and across many instruments, and I have seen calibrations predict over years and across instruments but not good enough for ISO15197. Also S/N -as measured by precision- I have seen as low as 7-10mg/dL, hardly poor signal to noise ratio. But bias is another matter. Raman is a weak signal, so yes it may need to be continuous, but it has a molecular "fingerprint" for the best selectivity for glucose. I find it interesting that many who are on the outside looking in don't understand the real issues in the various methods. Its hard to be a guy like John Smith(the deceitful turkey) who was never involved with a significant success in noninvasive arena look at other tech and try to understand it all in such a secretive field. I have been involved in 2 companies in the past that had degrees of success in measuring glucose noninvasively using optical methods. If I had not seen the data, I would never have been able to guess the technical limitations from the outside and optical spectroscopy is my field. NIR has an instrument size issue, not a S/N issue. Downsizing may eventually come as technology advances, but larger instruments are a hard sell in this market. the encouraging part of the google effort is the well understood tech, much is known about enzyme electrodes. the discouraging part is perhaps the worst sample lag in all body fluids, and high sensitivity required due to the low concentrations. Attempts to correct lag without meal information have been tried for many years by some very skilled people in academics and industry, and I don't think there is any new math that suddenly promises an accurate correction there. By the way Mark, good luck with the RF, another 5000 data points could shift those who don't believe but it may take that many or even more.

We think our device that is 10 x as thick as an optical contact lens and sits entirely under the lid, avoiding the problems listed above regarding vision, dryness and irritation, offers more volume to serve as a platform for drug delivery or device options:
http://www.youtube.com/watch?v=fzhXsvludLc
However we have worked only in drug delivery so have not addressed any of the issues above related to glucose monitoring in tears, which appears to be quite a challenge. But this set of hurdles can be isolated from those of a developing a multifunctional contact lens, which has had intuitive appeal for 40 years, but has never worked out as mentioned in comments.

thomas blank Mark, away from the marketing rant there, I appreciate the information. As technology moves forward I suspect there will be multiple noninvasive devices, especially now as investment moves out of VC realm. As for the sentiment, thanks but I am not working in this arena, just a spectator.

Check out EyeSense- a spin-off from Ciba Vision 8-10 years ago. So nothing new here !!

http://www.eyesense.com/en/ueber.htm

Mark Iverson Tom, Ms. Cusano, and all other readers of this Discussion...
I realize my friendly sparring and sharing of experiences with Tom about two different technologies used for the same purpose has somewhat diverted the subject of this Discussion, and for that, I apologize; had I received even one request to 'take it elsewhere', I certainly would have.

There are people in this Group that are on the investment side and I hope my interaction with Tom has provided them with a better understanding of both technologies to help guide their questions when evaluating some startup's claims.

Finally, all Google has to do is mention something and every news outlet covers it; the small startups are ignored, so how do they get the word out about promising technologies they're working on? I have only used this LinkedIn group to spread the word about our tech once or twice last year. Thank you for your patience and understanding. A healthy and prosperous 2014 to all Medical Devices members... -mark

Walt Bassett John L has reiterated what I said, but with a reference from Ciba Vision. This is old technology and one with a technical push rather that satisfying a customer need.
There are simpler ways than a contact lens for determining Blood Sugar.
and if you're thinking drug delivery..... also old news, but only useful for very specific drugs that have to go through that short pathway.
Suggestion for the scientists here..... avoid sticking ones head into the technical sand, and look around for a need that isn't already being satisfied already. there are dozens of analytical ways to measure sugar, but are they simple and inexpensive? If not, they are just novel solutions with no business case.
What does a contact lens method provide that is better than any existing method available today? Cost? Ease? Accuracy? OR....perhaps I am jaded for having worked on this one before ;-)

thomas blank Mark,
It is difficult for me to read speculation when I know more facts about the technology that I have worked on. Fundraising/timing ultimately and killed both companies that I worked at, not the technology. that is not to say that limitations in these technologies don't exist they do. But most observers don't know what they are. I sympathize with the fund raising challenges for a startup in noninvasive glucose. I would rather see the startup win, its a tough environment. I probably shouldn't have used the word "luck", I know the technical part is not luck, its the financing where luck may be needed. Best wishes for success!!

Carol David Daniel There are no convenient, non-invasive, real time, portable and cost effective methods of accurately measuring blood glucose on the market that would be suitable for self monitoring by a diabetic, so there is no existing method to make a comparison with.

Willi Glettig Dear Walt
A lifely and interesting discussion here!
I am in the molecule separation business and I don’t know any analytical technique that sees only or can single out glucose molecules from a heterogeneous, continuously changing mix of other molecules.
To separate glucose from the rest and to quantify it is also a challenge.
Other strategies e.g. specific modification of the glucose molecule so as to make it unique and detectable with a low cost technique is currently purely speculative.

Maybe Google should not waste their money on reinventing the diagnostic wheel (the currently used invasive techniques seem to work fairly successful). Maybe they should improve therapy (other forms of applying insulin or new drug discovery)?

Google being a pioneering company should do everything to stay a pioneering company. If Google aims to help peoples, than they should not try to compete with current diagnostics, pharmaceutical or medtech developments. Maybe they should finish of what the pharmaceutical industry has abandoned some 10 years ago that is discovery and development of new antibiotics.

Walt Bassett Perhaps Google might find a whole new application, since they are not bound by conventional pharma wisdom. I hope they just learn from the available prior art and not go down redundant rabbit holes.

Bogdan Baudis I agree with Walt. This project should be thought about in the terms of an platform rather than a specific application. Choosing the hard application may not be a bad idea to motivate the team and test the boundary capabilities. Nothing says that it has to be only a glucose sensor ... Also (sadly) the US market probably is NOT going to be the one where this platform is going to be deployed first ...

Having worked in continuous glucose monitoring for almost two decades, my interest was initially piqued by Google's announcement, until I did a little digging on the technology they're using (conventional electrochemical glucose oxidase/hydrogen peroxide detection). Besides the fact, as many have commented above, that the concentration of glucose in tears doesn't seem to track that in blood very well, there are issues in using the sensor that they seem to be using with an "episodic" measurement, that is, a measurement that is triggered whenever the user hold the reader up to the contact lens. Maybe they've changed their operating paradigm since the last UWash paper was published, as there weren't many details in their press push.

In summary, interesting engineering, but not a real useful product as I'm able to determine from the available technical info.

Mark Iverson Tom:
Much appreciate the additional details on the various 'flavors' of optical spectroscopy... and the various advantages and disadvantages of each. Perhaps a bowl of several different flavors could overcome all the limitations to achieve what's necessary for a marketable device -- but it'd probably be a device even Rube Goldberg would be proud of! :-)

RE: your comment,
"The RF method sounds interesting, though there is a question whether it is specific to glucose or whether water shifts or some other analytes impact the signal. A diabetic on diuretics, as many late stage insulin dependent diabetics are, will experience large water shifts not related to glucose."
I would bet our SNR and repeatability are much better than any optical spec given a 5 second scan, and you are correct that specificity has been the 'nut to crack' for RF, but I think we've done it... but it can't be done using a narrow band of the spectrum. What one finds is that if you look over a broad enough span, and combine the information content properly, you can achieve specificity for different analytes.

RE: the information content within the acquired signal...
An additional advantage of RF measurements is that they contain much more information than optical. We can look at transmitted AND reflected simultaneously, amplitude AND phase, resistive VS reactive component, etc. Wavelengths of optical methods are just too small to give you any kind of phase information, so you're mostly limited to amplitude only. I just came across this paper today from CalTech that you might find interesting:
irmmw-thz2014.org/sites/default/files/IRMMW-THz2014_Abstract_Example.pdf

Thanks for the 'good luck' wishes, and if you were going to invest in a technology, would you bet on the tech that has been tried for 30 years and collected *millions* of samples and still has NO marketable device to show for it, or the newcomer whose 800+ samples have already achieved predictive accuracy of several weeks with no recalibrations? I'm bettin' on the newcomer! ;-) And I haven't even begun to list the business advantages of the tech...

B Well and best of luck in your efforts too... hopefully someone will succeed soon!
The diabetic community has been waiting way too long for a noninvasive glucometer.
-mark

Mark Iverson I've talked with a few friends who wear contacts and they wonder how long it takes to reestablish accuracy after using lubricating eye-drops!

Frank Jurik Michael T. and Steve Z.- Also having worked for over two decades in the area of glucose measurement (episodic and continuous) I studied tear glucose measurement. In 2007 Dr. Sanford Asher of the U. of Pittsburg published a nice review of the past history and challenges inherent in determining a possible correlation between tear and blood glucose.
see: http://www.pitt.edu/~asher/homepage/colloid_pdf/Tear+Glucose+Analysis+for+the.pdf
As far as I know, a reasonable correlation has not been obtained. It will be difficult as tear glucose concentration is very low, (i.e. at a fasting blood glucose level of ~80mg/dl the tear glucose is only about ~0.5 mg/dl). In addition, it appears that any slight perturbation in sampling technique causes large variation in tear glucose. Additionally, the transport of glucose from the blood to the tears is not well understood.
In summary, this is a very interesting and challenging project that could keep some scientists employed for a number of years. In my opinion, it is very high risk but (as others have already mentioned) might be a platform for other spin-off technologies. Finally, I believe companies like Google should be doing basic research whether or not the final product is obvious. However, not so certain about the news splash aspects of the work. There is lengthy history of raising and dashing the hopes of diabetics with non-invasive news releases. Hopefully by now the novelty has worn off.......

David Lim, Ph.D. RAC, CQA I hope Google can find a way to crack the mystery of measuring glucose in tears with much higher accuracy than what is currently available using blood although it is arguably skeptical. It seems Google is currently working on prototyping with some tests done already with promising results. But there are lots of hurdles to overcome to finally bring it to the market. As long as glucose detection in tears can be made in a SUPER-sensitive manner, it should be achievable.

Juan José Janer Don't you thiink we are having just a little bit of information about the google lenses, but, enough to maintain this kind or debates?.
I blieve that there are many ways to observe and amplify the data collected from the vascular corneal blood stream (no tears needed) and work with algorithms in order to detect posible low glucose levels problems and or send special message to the patient?.

Matthew Romey As Thomas and many others have mentioned, it seems a long shot that measuring glucose in tears will meet the accuracy requirements for the FDA *or* for the customer.

Here is a nice review article; a little dated but interesting nonetheless: http://www.ncbi.nlm.nih.gov/pubmed/17938838.

"While several methods for sensing glucose in tear fluid have been proposed, controversy remains as to the precise concentrations of tear glucose in normal and diabetic subjects and as to whether tear fluid glucose concentrations correlate with blood glucose concentrations."

This subject along with reading peoples responses has attracted and kept my interest. I have two teenages that are Type 1 on insulin pumps. My oldest was diagnosed when she was 13 mos. old and my other was diagnosed at 10 yrs. old. My oldest currently wears daily disposable contacts which also has my interest. My family is active with the JDRF and ADA donating time and money to help find a cure. Even though Google contacts is not a cure, I am interested that science and technology is "thinking outside the box" to find ways to make living with diabetes easier and less costly. Keep up the good work!!!

David Lim, Ph.D. RAC, CQA I envision Google's technology may be related to this technology (a biosensor for the detection of glucose and other components in tears) at http://wp.me/P4i0FB-gd

David Austin PhD Instead of a contact lens, if tears can really provide the necessary results data why not have a system that can remove a sample of tears which is then analyzed?

David Lim, Ph.D. RAC, CQA @Dr. Austin, that could be an option. However, if tears are analyzed in situ, data could/would be more accurate/reliable and less variable. If you do so, you might/could overcome the prior art, arguably speaking.

David Austin PhD Dr Lim, Thank you for your comments supporting a more practical route for monitoring, as I said some time ago along with others, diabetics and contact lenses present a higher risk group when compared with non diabetics wearing contact lenses especially if worn for overnight use.

Mark Iverson Mr. gianni sardi:
Please post your in-vivo results so we can see how they compare! Our data is in a PDF file on my profile page... Have you achieved predictive accuracy over time... what are the statistical results... how many paired samples... Enquiring minds want to know! And if you're open to combining technologies, please get in touch to discuss.

Mark Iverson Drs. Lim and Austin:
If you haven't, you might want to read all the comments in this discussion. In Frank Jurik's response, he posted a link to a paper on the use of tear fluid for glucose measurement. The use of tear fluid, whether in-situ or not, has very significant issues which make it one of the less likely methods to achieve the necessary sensitivity and repeatability. E.g., many contact lens wearers use eye-drops to lubricate their eyes... which dilutes the glucose to unbelievably low levels. How long must they wait before it has recovered? There is also a very significant lag-time. Fortunately, the media attention about Google has sparked renewed interest, so things are looking up.

Mark Iverson Bernard,
That is the same John Smith, PhD, who wrote a very lengthy summary of the efforts to do noninvasive glucose. You should be able to find it on the web as a free downloadable PDF. It's called "The Pursuit of Noninvasive Glucose" and he recently updated it to the 3rd edition. It is highly recommended for anyone wanting to understand what's been tried, and the caveats and errors made by past efforts. Noninvasive glucose is a difficult measurement to make. I've been working on and off with the RF-based technology since 1993. In 2008, our angel took control of a poorly run company and we made very good progress. In 2010 we finally achieved predictive accuracy over weeks/months, with no recalibrations.

David Austin PhD Many thanks for your earlier reply, my interest has always been in more direct methods for glucose measurement. however, I do have clinical skills in eye research and contact lenses so always feel obliged to comment on practicalities about the application of contact lenses as a medical diagnostic device.

Joel Blatt Biological variability will trump technological hubris any day -- as any of the experienced IVD folks above have testified. Also, why do something technologically slick when an established technology will work just fine? Interference issues can be resolved with established IVD technology and small volume blood samples can be acquired readily. There is a future in non-invasive (Raman?) but there also has to be a compelling economic story.

Bernard Farrell There's an interesting blog post on DiabetesMine by John Smith, it's from 2009 but relevant to this discussion.

Quoting: "Attempts to measure glucose in the eye, where transparent structures make viewing easier, are among some of the oldest and most frequently investigated."

http://www.diabetesmine.com/2009/08/the-search-for-noninvasive-glucose-technology-that-works-where-it-stands-now.html

Mark Iverson Joel Blatt wrote:
"why do something technologically slick when an established technology will work just fine?"
Do you realize how absurd that statement is when it comes to noninvasive glucose monitoring? It has nothing to do with 'slick technology', and everything to do with making life less painful, safer, and healthier for the ~380 million people who have diabetes. You obviously are not diabetic, nor been around them for any significant time to understand what they go through on a daily basis...

RE: "There is a future in non-invasive but there also has to be a compelling economic story."
Ours is actually quite a compelling 'economic story'...

Benjamin Richard Wang MD Does anyone know the physiological reason why patients with Diabetes develop Macular Degeneration?

This is the same reason why saliva and urine will not work in this field. The relationship between blood glucose and tear glucose levels is related to the physiological non-insulin dependent transport of glucose into ocular tissues. The osmosis of water dilutes the concentration of glucose, but the ocular pressure of the vitreous fluid chambers of the eye change accordingly.

Google probably knows that, I can't imagine they don't have a smart physician helping them. The key to measuring blood glucose from vitreous fluid is not to measure glucose in tears. The key is to track the pressure changes inside the eye, which can be done easily with a contact lens, we can already do it with simple devices, used to screen for glaucoma. This is not a linear or even a simple polynomial relationship, but it can definitely be done.

Yes Google can do it and will do it because unlike many other of the other dozen companies trying to measure saliva glucose, because they are big, smart, and most importantly, they are chasing the right fluid. They will find the right approach.

David Austin PhD Dr Wang,
The relationship between diabetes and MD is a well understood cascade related to vasular change, exudates, nutrient reduction and hypoxia. This is why in patients with poor control undergo laser to reduce these effects along with anti-VEGF drugs etc.
I think you are referring to aqueous fluid (not vitreous) which is supplied by blood which is modified by the ciliary body part of the uveal system. One route for glaucoma is the incorrect level of production of this process.
Tears are supplied primarily by extraocular glands.

Matthew Romey Dr. Wang: Are you saying that changes in blood glucose can be detected by measuring intraoccular pressure? I've never heard of this and have never seen any literature to support it. Can you provide a literature reference?

David Austin PhD I agree, that does not make any sense at all. The aqueous is however altered with respect to glucose levels when a diabetic maintains poor control this is what affects the lens and induces prescription changes. Is this the point of confusion?

David Lim, Ph.D. RAC, CQA Dr. Want may be referring to similar technologies (blood sugar, pressure in blood vessels, intraocular pressure measurement, etc.) at http://wp.me/P4i0FB-hd

Mark Iverson I too have not heard/read anything about intra-ocular pressure correlating with glucose levels, however, it does make sense for the longer term. I think it is common knowledge for those involved in diabetes research that labile glucose levels cause damage to the microvasculature and peripheral nerves. I've read several papers about how capillary membranes become 'leaky' as a consequence of not only diabetes, but several other maladies. This leakiness could very well be the cause of the increase in intra-ocular pressure.

One other application for our noninvasive tech is a very low cost device which can measure the degradation of capillary permeability -- i.e., diabetic screening. We could do that 15 years ago, but couldn't convince anyone that it was of any use! I just saw a study that concluded that 46% of the people with diabetes (globally), are unaware of it. That's ~170M people who will probably suffer significant tissue/nerve damage which could have easily and inexpensively been detected before it did the damage. We're workin' on it!

David Lim, Ph.D. RAC, CQA @Mr. Iverson, have you taken any freedom to present your research plan to the FDA as part of Pre-IDE, Pre-Submission (now called Q-submission or Q-sub)? I presented yesterday. Here is some info for your consideration if you haven't done done so. http://wp.me/p33LGu-2ad

Mark Iverson Dr. Lim:
Thank you for taking time to make that suggestion...
Not at this time, but it's been discussed. My partner is on the East coast, and has spent a lot of time in med-devices and working with the FDA. Like many startups, we are somewhat 'resource limited'...

Alan Rapacki I worked with Foresight Labs Lab in Menlo Park, an incubator devoted to novel therapies for treating eye disease. I went on to work with QLT developing an implantable punctum plug to treat glaucoma. On thing we learned is that the tear film as well as the eye is often a leading indicator for other diseases. I believe that it is possible that IOP can be a predicator for other conditions. Although I can accept that this is possible, I not sure if Google has the expertise to gain FDA approval for this device - I have been there myself and not sure if simply the "Google Name" is enough to can get them through the PMA, or 510K process and on to commercialization.

Parents of a Type 1 infant would gladly swap a non-invasive method to test their son or daughter over a finger prick 4 times a day. Much like the glass thermometer has been replaced by a digital version, there will be advances if we embrace "better, faster, smarter". And to address the economics, how many more incremental disposable strips will be sold if the measurement is less intrusive and easier?

David Austin PhD I completely agree if there was a method to collect a tear film sample without the need for a contact lens.
However if a daily disposable format was possible for those needing a vision correction that seems another possibility.

Ete Szuts Because tear is filtered serum, its glucose levels lag glucose changes in blood by 15 minutes or more. But it is BLOOD glucose, not TEAR glucose that needs to be controlled for good diabetes management. And current medical practice is pushing for greater not lesser accuracy in blood glucose measurements. Hence, tear measurements will never be as therapeutically useful as blood testing. In its limitations, tear testing is similar to and as useful as urine testing -- the old antiquated assay that was superseded decades ago by blood testing. FDA is unlikely to approve any new diagnostic test for blood glucose control that is not better than predicate devices.

Alan Rapacki Hi Ete, you make a great point, the standard of care is blood testing. It is an accurate test with a well established systems in hospitals and private labs. It will be difficult to commercialize this device if it is not as accurate as the current method - the risk of a wrong diagnosis is too great.

David Austin PhD Surely a 15 minute time lag is a minimal concern, this could surely be factored into their schedule. My concern is that wearing a contact lens for this at risk group with respect to wound healing and recovery is not the best way to go. However, if there is a more straightforward test on control then that has got to be preferable. It is interesting that there is also a blog on why medical devices fail currently being discussed on linkedin, often the reason is need and simplicity for the patient.

Ete Szuts Actually a 15 min lag time is of major concern when blood glucose is changing rapidly. A >10mg/dl change in 1 minute is not unusual under certain circumstances (see publications by others or me). But lag time is not the only limitation. There are other physiological sources of error: availability of glucose for filtration due to blood flow fluctuation within integument (flow in superficial capillaries varies greatly, can even stop, while blood glucose concentration remains unchanged flowing within deeper vessels), rate of tear production by gland, breakdown of glucose during filtration, to list some of the more significant ones. Effect of these are hard to control and measure -- and are also variable between people. When the errors from these are compounded with errors intrinsic to the measurement device, meeting ISO requirements for accuracy and precision (95% of device tests need to be within 20% of blood glucose by YSI gold standard or equivalent) is very hard to achieve for the duration of product use. Above is based on my many years of having worked with glucose diagnostic devices (both in vitro and in vivo, whole blood and interstitial, etc.).

Matthew Romey Ete: Your post is accurate and, in fact, I have already expressed similar concerns in this discussion over the physiological and technological challenges to accuracy.

Assumedly, the excitement about this product concept stems from the desire to measure glucose without the need for frequent blood collections via fingersticks. Unfortunately, for reasons already mentioned, a contact lens or non-invasive product will be unlikely to have the accuracy to *replace* direct blood measurements.

However...

It could provide other benefits that intermittent fingersticks do not. Continuous tracking is one. Fingerstick measurements provide no information about direction of change. Say you are at 80 mg/dL... are you going up or down? That's really important information that mitigates the need for absolute point-to-point accuracy in line with updated ISO and FDA standards for glucose meters. Linkage to wireless health is another - Dexcom and other CGM manufactures are already doing this.

Extending this thought, one could envision a device that provides the tracking without a claim to absolute accuracy. This might be done in numerous ways: preclude putting a number on the screen, perhaps replaced it by a shade or a color; or limit the resolution of the number to, say, 5 or 10 mg/dL. Alarms could play a crucial role.

My point is, there is valuable information in continuous tracking even if the accuracy is not perfect, as long as it is an clearly indicated as an adjunct to blood sample measurements with reliable meters.

Jerry Robinson This is one part of a technology product that will save lives.. probably a lot of them... continuous monitor will directly benefit people.

David Austin PhD This is what I was saying in response to Ets Szuts , the time frames that people take their diagnostic values using a prick test will surely be a variable so, I still do not see why the lag is that important as long as the value obtained is representative of control.

Ete Szuts The benefits of continuous glucose monitoring have been well known for diabetes professionals for decades, those being trend tracking, alerting wearer of hypo conditions, detection of nocturnal hypoglycemia, semi-invasiveness, etc. Of course, all that is desirable.

But that was not the point of Ms Cusano's posed question.

To get to market, Google's contact lens needs to meet performance of predicate continuous glucose monitoring device approved by FDA: Medtronics CGM System (Guardian and iPro2), which measures glucose in interstitial fluid with a subcu sensor that must be recalibrated with fingerstick at least 2x per day and even after that does not display glucose results to users in real time (they get it many days later from their doctor). Google's device is unlikely to track blood glucose as well as Medtronic's CGM, because Google does not measure interstitial glucose, it measures tear glucose and tear is modified interstitial fluid, processed by tear gland. Google's device is being burdened by more physiological variability.

Frank Jurik Ete- Note that the 7 day Continuous Monitors from Dexcom have FDA approval and are CE marked. They display glucose trends in real time and have a hypo alert. However, they must also be re-calibrated every 12hrs and any treatment decisions must be based on Fingerstick data and not on the CGM values.

Mark Iverson Ete comments on the Medtronic CGM technology,
"must be recalibrated with fingerstick at least 2x per day"
and Frank comments on Dexcom's competing product,
"must also be re-calibrated every 12hrs..."
BOTH of which are invasive (implantable)... AND replacement sensors are costly.

Our noninvasive technology remained accurate for 4 weeks after calibration, with NO recalibrations. Calibration is not yet universal, but some initial efforts were encouraging that a universal algorithm was possible. The first company that approached us did not want to commit to furthering development, so we are continuing to pursue other ways to move this forward. Any suggestions or referrals would be greatly appreciated...

Juan José Janer I have made several comments about the goal of improving measurement systems that are useful for both the carrier device and for medical services or specialists.
The half-life of insulin in the body produced by the beta cells is approximately + -. 5 minutes, the sensitivity and specificity of a continuous monitoring system to be balanced is very complex.
The half-life of insulin in the body, produced by Beta cells, is approximately + -. 5 minutes, the sensitivity and specificity of a continuous monitoring system to be balanced is very complicated, but hardly impossible. They have the money, the capacity and the open market.

Since 2004 and in strong R&D cooperation with leading universities, we developed a non-invasive and continous (every 6sec) blodd glucose measurement device. After preclinical and successfully execution of clinical trial, a patent was granted. Now we chance the focus of our developments to commercial applications, e.g. "Wearables" and "Digital Health".
By preparing the R&D project and by project execution and by execution of trials, we got a lot of experiences in measurement of blood glucose, its physical and medical parameters, etc.. Especially the sensors, the use of intelligent SW and HW filtering and the use of artifical intelligence was neccessary for processing and visualization the non-linear and individual function(!) of blood glucose. I believe, that wide ranges of experiences, good ideas, enormous engagement, positive thinking and persistence shall be necessary.
I also agree with some comments here, that they have the money and thus the access to almost unlimeted resources and much time, of course.

Paul M. Stein Helmut, you very astutely point out the crazy non-linear, multi-compartment nature of glucose transport within the body. That is a huge nut to crack with regards to tight maintenance of blood glucose control with insulin administration; i.e., the artificial pancreas. Best wishes to you. Apple may not have the wherewithal or the stomach to handle such an intensive undertaking besides what they have right now.

Lou Cabana While I find the technology interesting, there are a number of hurdles Google will have to overcome. The new standards for accuracy and precision that both the FDA and the EU are implementing are going to be a challenge even for current SMBG mfgs. Added to that the need to be equivalent to existing devices for interference and temperature performance and you have a minefield to navigate. Although the market place is less attractive( lower reimbursement rates and more than 1 of the big 4 players are for sale) you have a interesting landscape to navigate. i would not bet against them, but steep price to be paid to gain entry into this space.Agree with Paul's comments. Having spend last 15 years in this space, the technical and business challenges are formidable.