Graphene tattoo: the electronic tattoo enables continuous blood-pressure monitoring using electric currents and bioimpedance measurements. (Courtesy: The University of Texas at Austin/Texas A&M University)
Wrapping a cuff around a patient’s arm and inflating it to measure blood pressure is one of the most routinely performed medical tests. It provides a quick and reliable assessment of cardiovascular health, as blood pressure is an independent predictor of all-cause mortality. But such arm cuffs are bulky and uncomfortable, making them impractical for continuous monitoring outside of clinics.
For this reason, researchers are developing cuffless alternatives with the goal of unlocking new possibilities for patient diagnostics and management, as well as providing new understanding of physiology. However, none of these tools has become a mainstay yet.
One option, acoustic sensors, slide during movements and are too large to be easily incorporated into untethered ambulatory sensors. Meanwhile, optical modalities such as smart watches are limited by the low penetration of light into tissues, which hinders their ability to capture haemodynamic parameters in the arteries. Studies also show that optical sensors are sometimes inaccurate when used with darker skin tones or larger wrists.
Graphene tattoos and machine learning
A group of researchers from the University of Texas and Texas A&M University, led by Roozbeh Jafari and Deji Akinwande, circumvented these impediments by developing a sticky and stretchable graphene electronic tattoo that is comfortable to wear for long periods and does not slide around. They describe the new blood-pressure monitor in Nature Nanotechnology.
Near-invisible: with a 25 mm2 surface area and a thickness roughly 1000× thinner than acoustic sensors for blood-pressure monitoring, the graphene electronic tattoo attaches seamlessly to the patient’s skin. (Courtesy: Roozbeh Jafari)
Graphene, one of the strongest and thinnest materials in existence, is similar to the graphite found in pencils but with the carbon atoms precisely arranged into layers just one atom thick.
“The sensor for the tattoo is weightless and unobtrusive. You place it there. You don’t even see it, and it doesn’t move,” says Jafari.
The device performs measurements by injecting a low-intensity electrical current into the skin and then analysing the body’s response, known as the bioimpedance. The electrical signal penetrates deep into the skin and propagates through the path of least resistance: the blood vessel, as blood is ion-rich and thus a better conductor than the surrounding fat and muscle cells. The signal that is collected reveals variations in bioimpedance, which are correlated with blood-pressure variations.
The researchers also used the device to measure pulse wave velocity, the speed at which blood travels in the arteries. They then used the bioimpedance and pulse wave velocity data as inputs for a common machine learning algorithm (AdaBoost) to predict diastolic (minimum) and systolic (maximal) blood pressure points.
Grade-A classification performance
To assess the accuracy of the tattoo, the researchers enrolled seven volunteers, attached sensors above their radial arteries and asked them to perform a series of activities known to change blood pressure (hand grip, cycling on a stationary bike and the Valsalva breathing manoeuvre). They measured reference blood pressures using a medical-grade blood pressure cuff.
In total, the researchers recorded 18,667 data points and split the data into 89% for training and 11% for testing, a process known as cross validation. The measurement accuracy – 0.2±4.5 mm Hg for diastolic pressures and 0.2±5.8 mm Hg for systolic pressures – was equivalent to grade-A classification, according to the IEEE standard for blood-pressure monitoring devices.
Read more
Wireless device eases blood-pressure monitoring for children in intensive care
Further, some activities were combined with sweat-inducing walks outside at 38 °C or push-ups; none of the sensors degraded electrically after exposure to light and heat or contact with water or sweat. The sensor was able to monitor arterial blood pressure for more than 300 min, 10-fold longer than reported in previous studies. The sensors could also be used to record other vital signals, such as breathing respiratory rates, and could be placed on other locations, such as the tibial and carotid arteries and the jugular notch.
Looking forward, there is still some work needed to estimate central aortic blood pressure, which differs from peripheral blood pressure and is thought to be a superior indicator of cardiovascular events. Likewise, assessing the whole pressure wave over a cardiac cycle as opposed to single points could provide additional information on blood vessel functions and cardiac performances.
