What Can Skin—and Sweat—Really Tell Us?

By Paul Nicolaus

October 7, 2019 | What can the body's largest organ and its secretions teach us about human health? While the answer may be a bit unclear for now, scientists are pursuing wearable skin sensors that can monitor bodily signals of interest and glean useful information from perspiration.

The hope is that these efforts will lead to new applications in exercise physiology and new potential for preventive care, diagnosis, and treatment.

Researchers based at Stanford University, Nanyang Technological University, and the Samsung Advanced Institute of Technology have developed a group of networked sensors, for example, that can wirelessly monitor physiological signals.

The technology, described in a paper published August 15 in Nature Electronics (doi: 10.1038/s41928-019-0286-2), includes two main parts. The first consists of soft, stretchable sensors that can be printed in various shapes and attached to the skin. The other component is a flexible readout circuit that attaches to clothing or sits in a pocket and sends signals from the sensors to a smartphone.

Over several years, a 14-person team developed the sensors. The intent was to come up with technology that would be comfortable to wear, and to accomplish that, the group avoided the use of silicon chips and batteries within the soft, on-skin component.

The main design novelty, lead author and Stanford University postdoctoral research fellow Simiao Niu told Bio-IT World, is the use of wireless communication to link the soft and rigid parts rather than using a traditional wired connection. To pull it off, the researchers revised radiofrequency identification (RFID), a form of technology more typically used for keyless entry into locked doors or buildings.

To demonstrate the concept, the researchers put sensors on the wrist and abdomen of a test subject to monitor pulse and breath. Additional sensors were attached to the elbow and legs for body movement detection. Monitoring pulse and respiration rate at the same time can be used to analyze sleeping conditions, and there are potential exercise applications as well.

Next Generation Work

The next steps will involve the integration of the wireless RFID platform with newly developed sensors, according to Niu, along with efforts to shrink the sensors to enhance comfort.

"We still want it as thin as possible and as small as possible," he said, "so that is the other part that we are trying to improve." Beyond that, the readout circuit developed is a prototype, and the intent is to reduce its size by integrating all chips into one and using smaller batteries to power it.

"The first target for us is for the clinical application," Niu said. Patients could wear this after surgery, for example, or to monitor and address an issue like sleep apnea. There are aspirations to apply this technology to consumer electronics, too. "We want to develop robust sensors that you can wear anytime, anywhere to monitor your physiological signals," he added.

The vision is that the collection of wireless sensors could work along with smart clothing to track a greater variety of health indicators than the smartphones or watches currently worn by consumers. "We think one day it will be possible to create a full-body skin-sensor array to collect physiological data without interfering with a person's normal behavior," Stanford University chemical engineering professor Zhenan Bao said in a news release.

For future development, additional computing power could be added to the sensor node, University of California San Diego researchers noted in a commentary piece published in Nature Electronics (doi: 10.1038/s41928-019-0291-5). While difficult to pull off, distributed computing is important for ensuring long-term viability, particularly when used in the realm of personalized healthcare.

Beyond that, the incorporation of optical, chemical, and biological sensors could expand the impact of the platform. Although there's still work to be done, Sheng Xu and coauthors referred to this sensor network as a step toward the next generation of wearable electronics and the Internet of Medical Things.

There are currently efforts underway, Niu said, to add more sensors to the technology platform to monitor aspects like body temperature or perspiration. Sweat contains electrolytes and cortisol, for example, which could provide insights into areas of interest such as dehydration or stress.

This complex make-up, along with its ease of access, means sweat could be a useful means of better understanding the human body at a chemical level. The big question, though, is what sort of useful information can be gleaned from this biofluid?

Frustrations Posed by Perspiration

Sweat analysis has been used extensively in the past for some applications, Ali Javey, professor of electrical engineering and computer science at the University of California Berkeley and faculty scientist at Lawrence Berkeley National Laboratory told Bio-IT World.

The gold standard for detecting cystic fibrosis for the last several decades has been sweat analysis, he said, "so we know there are some cases where sweat has been known to be very relevant."

Javey and colleagues want to look at this bodily fluid from a broader perspective, though, to better understand whether sweat glucose, the concentration of ions, or the concentration of drugs (or drug byproducts) are informative.

To learn more about the composition of sweat, however, there is a need for reliable sensors that can be placed on various parts of the body on large numbers of subjects. And coming up with this type of tool means coming up against an array of hurdles.

On-body measurements are ideal because it enables the collection of discrete data points as well as continuous measurements. "You can actually look at dynamic changes, for example," he said. One problem with sweat analysis, though, and especially sweat analysis on the body, is the small sample volumes.

"We are often dealing with fairly low analyte concentrations—much, much lower than you would find in blood," he said. From an analytical chemistry point of view, "you know you're in trouble when you have very little sample and you have very low concentrations."

Evaporation is an issue of concern, and skin contamination can be problematic as well. Many soaps have sugar in them, for example, which could interfere with efforts to measure sweat glucose levels.

Yet another dilemma is finding ways to ensure that only the freshest perspiration is taken into account. "If you want to do dynamic studies," he said, "you want to make sure that you're only measuring the most recent sweat that is generated."

New Sensor Design Revealed

In a paper published August 16 in Science Advances (doi: 10.1126/sciadv.aaw9906), Javey and colleagues at the University of California Berkeley and the VTT Technical Research Center of Finland reveal a new sensor design that attempts to tackle these types of difficulties.

Reducing evaporation was one area of interest, and this is accomplished using a microfluidic channel. Sweat is essentially forced through the channel as it is produced, which pushes old sweat out of the way and allows the researchers to perform more accurate composition analysis as a function of time.

Another noteworthy aspect of their device, according to Javey, is its ability to measure sweat rate in addition to sweat composition. "If you're exercising and you're sweating too much, or if you're sweating a lot and all of a sudden you stop sweating while you're still exercising a lot, these are signs of a problem."

More importantly, he said, they have shown that it is possible to mass-produce these devices using a roll to roll printing scheme. "It's a very scalable process," he said, and there is little variation between the devices. "And that really was important to allow us to start to do subject studies."

The sweat sensor has two layers. There is one component built on a piece of plastic that is in some ways similar to a band-aid. The second component has the electronics port on it that contains a microcontroller and a Bluetooth chip to transmit the data wirelessly.

Chemical sensors can measure concentrations of electrolytes like sodium and potassium as well as metabolites like glucose, and a spiraling microscopic tube wicks sweat from the skin and tracks how fast it moves through the microfluidic to measure sweat rate.

Electrolyte balance is particularly relevant in the realm of athletics, and to take a closer look, the researchers put sweat sensors on various locations of volunteers' bodies to measure the sodium and potassium levels in their perspiration as well as their sweat rates as they rode a stationary bike.

Their preliminary results revealed a strong correlation between sweat rate and body weight loss. "That's really important," he said, "because the main definition of dehydration is 2% body weight loss." Local sweat monitoring could be used to estimate the body weight loss linked to sweating, and "that would be very useful for many athletic applications."

They also used the sensors to examine the sweat and blood glucose levels in healthy and diabetic subjects. There has been hope that non-invasive sweat tests could replace blood-based measurements for the monitoring and diagnosis of diabetes, but their results indicated that there isn't a simple correlation between sweat and blood glucose levels.

"While this finding limits the establishment of universal, sweat glucose-based thresholds for diabetes diagnostics," he and colleagues concluded in their paper, "it does not bar the possibility of more complex, individual-specific correlations existing between sweat and blood glucose."

Decoding Sweat

To gather additional insight about the possible uses—and limitations—of sweat sensing moving forward, the dynamics of other sweat analytes will be targeted in large-scale studies. In addition to the efforts to make the devices more robust to allow for these more extensive studies, there is work underway to explore the relationship between medication and perspiration.

"We have had multiple papers already looking at the concentration of different drugs in sweat," Javey explained. The concentration of L-Dopa, a drug that is used for treating Parkinson's patients, "seems to be very well correlated with the dose intake of the subject," for example. This could be promising, he said, in terms of monitoring metabolism rate for a particular drug.

Parkinson's is an interesting example because the dose of the drug is crucial. "If the dose is too low or too high, the patients often do not properly function," he said. "But the problem, of course, is they take a pill. There's an absorption time, so the concentration keeps going up in the body at first. Then there's the metabolism rate, so the concentration eventually decays."

If there were a way to monitor that metabolism rate, it would help with the long-term hope of achieving personalized medication for patients. "So we are pushing in that direction very hard for drug analysis using sweat," Javey added.

Devices can be used to locally induce sweating without any exercise, so the subject doesn't even have to move—an aspect that could be important for clinical applications when patients may not be able to sweat via exercise.

"Not everything is going to be informative," he explained. Part of the research is merely figuring out which analytes show promise and which do not.

But the real key, according to Javey, is what he refers to as the decoding of sweat. "We really have to have a better understanding of what sweat composition is telling us."

Paul Nicolaus is a freelance writer specializing in science, nature, and health. Learn more at www.nicolauswriting.com.