Studies show early steps toward passive wearable monitoring of glucose, hydration, pulse

By: Jonah Comstock | Jun 12, 2014        

Tags: | | | | |  |

BOEx-glucose_watch-fig3-croppedAs the makers of wearable sensors continue to push the boundaries of what those sensors can track, new research suggests that a device worn on the wrist could some day track hydration, pulse, and even blood glucose non-invasively.

In two papers published this month in Biomedical Optics Express, the journal of the Optical Society, researchers from around the world used a technique called the speckle effect, by which a low-intensity laser is shined on the skin and the pattern of reflection is used to deduce chemical happenings beneath the skin. In the case of the glucose test, a magnet was also used to alter the polarization state of the light and make the speckle patterns more detectable.

In the first paper, researchers used the technique on a nondiabetic man before and after he drank a large grape soda. They tracked his glucose at regular intervals both with a benchtop version of the optical sensor and a traditional fingertip glucometer. They found that 96 percent of the time, the readings were within 15 percent of one another. The main source of the errors was movement of the subject, something that will need to be corrected for before a commercial wristworn device could be practical. According to the paper, researchers believe a motion compensator can be added in the near future. 

MobiHealthNews reached out to DexCom CEO Terry Gregg to ask about the feasibility of optical sensors replacing implantable devices for continuous glucose monitoring. He said that the research seemed solid and promising, but was “a science project” — very far away from a buildable, scalable device. Gregg said the movement problem may be a more significant challenge than the paper seems to suggest. It is also unclear how cheaply a device like the one in the paper could be built.

“We can build a sensor on a benchtop that is factory calibrated that will last two weeks,” he said. “That doesn’t mean we can build 2 million of them. You have to consider the scalability.”

Researchers also applied similar methods to hydration tracking and pulse tracking. The pulse can already be optically tracked at the wrist, but the speckle method could potentially improve accuracy. The second study proved that, at least for pulse tracking, motion compensation is possible. The hydration tracking covered in the first paper did not have quantitative outcomes data, but showed generally that the technology could be used to track hydration trends.

According to an OSA press release, researchers estimate that a commercial product based on these studies could be out in two to three years, a projection that should be taken with a grain of salt. However, the researchers in the glucose paper presented some possible use cases that would not require perfect clinical accuracy.

“If we consider that 95.7 percent of our optical measurements are within the error range, there might be a big usefulness of our technique in the medical practice due to the fact that we present a very simple device that might be applied for diabetic diagnostics,” researchers wrote. “Also please note that our device does not have to be used for medical diagnostics but rather for the field of wellness in which the people that use the device are healthy (e.g. people that monitor their diet or sportsmen) and the device is aimed to provide only the trend of the glucose concentration in their blood stream.”

  • everydayfella

    Here’s the reality with noninvasive glucose monitoring, and mark my words:

    We will literally have stem-cell-based pancreas regeneration or repair before a useful noninvasive glucose monitor is successful in the marketplace.

  • http://blog.bernardfarrell.com/ Bernard Farrell

    I agree completely with @everydayfella:disqus. As a type 1 I have enough problems with currently blood glucose meter technology. Being within 15 percent means that a blood glucose reading of 100 could actually mean the real value was 85, at which stage you need to be wary of hypoglycemia. I’d recommend reading Hunting the Elusive Turkey (PDF) http://www.mendosa.com/noninvasive_glucose.pdf