June 23, 2016
Pricking a finger multiple times a day to monitor their blood glucose levels often proves overwhelming for patients with diabetes. Many simply won’t follow their physician’s recommendation to test so frequently, making it harder to manage their condition.
“It’s a lot of stress,” said Dae-Hyeong Kim, PhD, an associate professor in the School of Chemical and Biological Engineering at Seoul National University in Korea. “Many patients don’t want to do it every day so they will do it once a week or once a month, and that will aggravate their diabetes.”
Kim and his colleagues have been working to develop a noninvasive way to monitor blood glucose using a tiny wearable electronic biosensor that detects glucose levels in sweat. In fact, preliminary research demonstrated that their Band-Aid–sized device not only monitors sweat glucose, but also might be coupled with microneedles to deliver medication (Lee H et al. Nat Nanotechnol. doi:10.1038/nnano.2016.38[published online March 21, 2016]).
The device is part of a new generation of flexible, wearable biosensors being developed worldwide as potential clinical tools. The devices build on the successes of the now ubiquitous wearable fitness trackers that measure activity, heart rate, and sleep using mobile technology. But they go a step further integrating advances in nanotechnology and materials science to detect clinically meaningful metabolites and other compounds in sweat and other bodily fluids.
“Mobile devices allow us to do this type of on-body monitoring,” said Joseph Wang, DSc, director of the Center for Wearable Sensors at the University of California, San Diego (UCSD). “It started with fitness and sports, but it is moving to biomedical applications.”
NONINVASIVE GLUCOSE MONITORING
Kim’s group is one of many exploring wearable sweat-based biosensors to monitor blood glucose levels. Although there are minimally invasive continuous glucose monitors currently on the market, these devices require insertion of a sensor under the skin to measure interstitial glucose levels and wirelessly transmit this information to a pager-sized monitor (http://1.usa.gov/1WfP7rh).
Artificial pancreas systems that couple these minimally invasive glucose monitors with automated insulin pumps that dispense the appropriate doses of insulin are also in development (Hampton T. JAMA. 2014;311:2260-2261). Advanced versions are moving toward the pivotal trials necessary to gain US Food and Drug Administration approval (Kropff J and DeVries JH. Diabetes Technol Ther. 2016;18[suppl 2]:S253-S263).
“The artificial pancreas systems still rely on minimally invasive [glucose monitoring] technology,” said Wang. “And they still need to be validated with a finger stick,” Wang noted.
Many researchers, like Kim, are looking to develop next-generation, noninvasive sweat-based wearable glucose monitors by building on evidence that the glucose content in sweat can be a faithful indicator of blood glucose (Moyer J et al. Diabetes Technol Ther. 2012;14:398-402).
“You can remove the pain and stress [of needle sticks],” Kim said. “But at the same time, you have challenges that are different than blood-based glucose monitoring.”
Glucose levels in sweat are much lower than in the blood stream, so accurately measuring these levels is more difficult, Kim said. In addition, pH level, body temperature, and the amount of sweat can skew glucose measurements in sweat and must be adjusted for.
To overcome these challenges, Kim and his colleagues layered miniaturized electronic sensing systems for each of these variables onto their flexible adhesive patch. To test it, they adhered the patch to 2 healthy male volunteers and monitored their glucose levels using a mobile device application that wirelessly communicates with the glucose-sensing patch. The sweat glucose levels measured by the device matched measurements from a commercial glucose meter used on the men.
While the results are promising, Kim noted that there are still potential obstacles, including variations in the skin condition of individuals that could affect sweat collection.
“The beauty is it is noninvasive,” Wang said of sweat-based glucose monitoring. “But there are still questions about how clinically relevant [sweat glucose levels are].”
To circumvent this potential limitation, Wang and his colleagues have developed a tattoo-based glucose sensor that uses reverse iontophoretic extraction to draw glucose from the interstitial space without puncturing the skin (Bandodkar AJ et al. Anal Chem. 2015;87:394-398).
“We want to do noninvasive glucose monitoring and get the same information [as the continuous glucose monitors],” Wang said.
Several groups of researchers are working to develop tear-based glucose sensors (Ascaso FJ and Huerva V.Optom Vis Sci. 2016;93:426-434). For example, Novartis has licensed Google’s glucose sensing contact lens technology (http://bit.ly/1HJCJFP).
Preliminary data suggest that noninvasive glucose monitors might also be coupled with minimally invasive medication administration. For example, Kim and his colleagues tested whether metformin-loaded microneedles on their sweat-sensing patch could effectively deliver the drug to mice genetically engineered to have a diabetes-like condition.
“We want to administer the drug automatically and hopefully invisibly,” said Kim.
The patch was adhered to the shaved abdomen of the mice. Unlike the microneedle patches being studied to administer vaccines, which dissolve immediately when they contact the interstitial fluid, these microneedles were designed to dissolve when the patch is heated up. Kim explained that this would allow the patch to release insulin only when the insulin sensing data indicated a need. When the patches were heated, Kim and his colleagues showed that not only was insulin released, but the mice’s blood glucose decreased.
Clinical validation and commercialization are the next steps for Kim’s device.
“We need a lot of collaboration and cooperation with physicians,” Kim said.
SOLUTIONS IN SWEAT
Scientists are also looking for ways to develop continuous noninvasive monitors for a variety of other health indicators besides glucose. Because sweat is rich in a collection of other metabolites and electrolytes, it can serve as a powerful storyteller. For example, an imbalance in sweat’s sodium and chloride levels can help diagnose cystic fibrosis (Jadoon S et al. Int J Anal Chem. 2015;164974). But often the sweat samples are taken at a single point in time and are analyzed later using bulky equipment, explained Wei Gao, PhD, an engineer and postdoctoral fellow at the University of California, Berkley.
“It fails to provide a real-time profile in context,” Gao said.
The flexible wearable biosensors that Gao and his colleagues have developed allow continuous collection of information about the contents of sweat while individuals are moving (Gao W et al. Nature. 2016;529:509-514). The sweatband-sized device uses tiny electronic sensors layered on a plastic band worn on the wrist or head that simultaneously measures body temperature, metabolites including glucose and lactate, and electrolytes such as sodium and potassium. The data can be wirelessly transmitted to a mobile device like a cellphone, allowing near-instant analysis.
“Now we can collect data every second and provide a real-time profile of sweat,” Gao said.
Real-time monitoring could also be very important in substance abuse treatment or research. In May, the National Institute on Alcohol Abuse and Alcoholism awarded $200 000 to BACtrack, a San Francisco–based company, as part of a contest aimed at spurring the development of continuous blood alcohol monitors. BACtrack developed a discreet wristband monitor that detects alcohol escaping through the skin and sends the information via Bluetooth to a smartphone. A second-place award of $100 000 went to a Santa Barbara–based start-up called Milo for its wrist monitor, which also detects ethanol molecules using a skin sensor. (http://1.usa.gov/1X0APfC)
Currently, ankle bracelets are used by law enforcement to monitor alcohol use among individuals who have committed alcohol-related crimes. However, these devices only take readings every 30 minutes and are stigmatizing. More discreet devices created as part of the contest might also be useful for helping accurately assess alcohol use among individuals in treatment or those participating in clinical trials. The real-time data might allow clinicians to intervene quickly in the event of a relapse.
To facilitate such everyday use, developers are focusing on practical considerations. The electronics in Gao’s sweat sensor were designed to withstand 2 hours of intense outdoor exercise, he said. To make wearable sensors even more comfortable, Wang and his colleagues at UCSD pioneered a method in 2012 that screen-printed sensors on a temporary tattoo (Windmiller et al. Chem Commun. 2012;48:6794-6796). To make the devices more attractive and fun, some electrodes have been printed in a design like letters or a smiley face. While the technology is developing at a rapid rate, in part because of multidisciplinary collaborations and technology company support, the real-world and clinical applications of these tools are still being studied. Real-time detection of dehydration might be one application, Gao said. Already, some wearable dehydration sensors are being tested in clinical settings and the military (Visser C et al. Conf Proc IEEE Eng Med Biol Soc. 2015;1271-1274; http://1.usa.gov/206Ey9F).
The center at UCSD is interested in developing sensors that might detect dehydration, nutritional deficiencies, and other factors that can be “precursors for falls that come from frailty, balance problems, poor muscular function, and many other problems of aging,” said Kevin Patrick, MD, a professor of family and preventive medicine at UCSD and a collaborator of Wang’s. Such home-based monitoring might help “sort out who needs help and who seems to be doing okay,” he noted. Having this information in real time might allow a clinician to intervene before a fall occurs.
A Massachusetts-based biotech company, MC10, is marketing a business card–sized biosensor to researchers studying patients with chronic diseases (http://bit.ly/27kkPZI). The pharmaceutical company UCB is using the device to monitor patient movement in a study of patients being treated for Parkinson disease (http://bit.ly/1Tb6tVa).
Before these sensors can help physicians monitor their patients in real time, more clinical validation is needed to ensure their measurements correlate with blood levels of the measured compounds and the individual’s physical condition, according to Gao.
Most of the devices are really at the “proof-of-concept stage,” Patrick added. Much work lies ahead of clinical use, particularly for devices intended for older individuals or those living with chronic disease, cautioned Patrick.
“Usability will be key for these,” he said. “Can normal people figure out how to use them and use them easily?”
Another challenge will be helping physicians meaningfully use the volumes of data these devices create. Patrick noted that most physicians are already struggling to use data from health apps and fitness trackers in the limited time of an office visit. He predicted that technology companies would step into a “middle-man” role, distilling the data down to clinically important information.
“On balance, we’re still pretty early [in development] with almost all of these sensors,” Patrick said. “But the creativity and innovation applied to these issues by these engineers is truly amazing and will likely be transformative to many aspects of medicine.”