The IEEE website places cookies on your device in order to provide you with the best user experience. By using our website, you agree to the placement of these cookies. To learn more, please read our privacy policy.
This wearable device senses blood sugar and manages insulin accordingly
The Tandem insulin pump is not bigger than a mobile phone. It injects insulin into the skin under the command of the Control-IQ software, which receives blood glucose data from the Dexcom G6 sensor.
In some ways, this is a family story. Peter Kovatchev was a naval engineer who raised his son Boris as a problem solver and built a model ship with his granddaughter Anna. He also suffers from a type of diabetes, where the pancreas cannot make enough insulin. In order to control the glucose concentration in the blood, he had to inject insulin several times a day. The syringe he used was placed in a small metal box in our refrigerator. However, although he tried to give the right amount of insulin at the right time, his blood sugar control was poor. He died of diabetes-related complications in 2002.
Boris is now working on bioengineered alternatives to the pancreas; Anna is a writer and designer.
People who need insulin must walk a tightrope. Blood glucose concentration fluctuates drastically, especially affected by diet and exercise. If it falls too low, the person may faint; if it rises too high and stays elevated for too long, the person may faint. In order to avoid repeated episodes of hypoglycemia, patients usually have high blood sugar in the past, which exposes them to long-term complications such as nerve damage, blindness, and heart disease. And patients always have to pay close attention to their blood sugar levels, they measure it many times a day by piercing their fingers to get blood drops. This can easily become the most demanding treatment that patients are asked to perform on themselves.
No more: artificial pancreas finally appeared. This is a machine that can sense blood sugar changes and guide the pump to inject insulin more or less. This task can be compared with the way a thermostat connected to the HVAC system controls the temperature of the house. All commercial artificial pancreas systems are still "hybrid", which means that users need to estimate the carbohydrates in the meals they will consume to help the system control blood sugar. Nevertheless, the artificial pancreas is a victory for biotechnology.
This is also the victory of hope. We clearly remember one morning in late December 2005, when experts in diabetes technology and bioengineering gathered in the Lister Hill Auditorium of the National Institutes of Health in Bethesda, Maryland, USA. At that time, existing technology enabled diabetics to track blood sugar levels and use these readings to estimate the amount of insulin they needed. The question is how to eliminate human intervention from the equation. An outstanding scientist stepped onto the stage and explained that the biological glucose regulation mechanism is too complicated to be replicated artificially. Boris Kovacev and his colleagues disagreed. After 14 years of work, they were able to prove that the scientist was wrong.
This is another confirmation of Arthur Clark’s First Law: "When an outstanding but older scientist says something is possible, he is almost certainly right. When he says something is impossible , He is probably wrong."
In a healthy endocrine system, fasting blood glucose levels are about 80 to 100 milligrams per deciliter of blood. The entire blood supply of a typical adult contains 4 or 5 grams of sugar-roughly equivalent to the sugar in paper coffee bags provided by restaurants. Eating carbohydrates, whether it is pure sugar or starch, will cause blood sugar levels to rise. A normally functioning pancreas recognizes incoming sugar and secretes insulin, allowing the body's cells to absorb it, so that it can be used as energy or stored for future use. This process restores glucose levels to normal.
However, among patients with type 1 or type 2 diabetes who require insulin — nearly 8.5 million people in the United States alone — the pancreas either produces no insulin or too little, and the control process must be approximated by artificial methods.
In the early days, this approximation was very crude. In 1922, Canada isolated insulin for the first time and used it for diabetic patients; a few decades later, syringes were the main tool used to treat diabetes. Because the patients at the time could not directly measure blood sugar, they had to test their urine, and the trace amounts of sugar in it could only prove that blood sugar levels had risen to disturbingly high levels. It was not until 1970 that dynamic blood glucose testing became possible; in 1980, it became commercially available. The chemically treated test paper reacts with the glucose in a drop of blood, changing its color according to the glucose concentration. Ultimately, meters equipped with photodiodes and optical sensors are designed to read the strips more accurately.
The first improvement is in blood glucose measurement; the second is insulin injection. The first insulin pump had to be worn like a backpack, impractical for daily use, but it paved the way for all other intravenous glucose control designs that began to appear in the 1970s. The first commercialized "artificial pancreas" was a refrigerator-sized machine called Biostator, used in hospitals. However, its size and the method of injecting insulin directly into the vein prevented it from progressing beyond hospital experiments.
The original artificial pancreas, called Biostator, was used in hospitals around 1977. It delivers insulin and glucose directly into the veins and cannot be adapted for family use. William Clark/University of Virginia
That decade also witnessed the work of more advanced insulin delivery tools: pumps that can continuously inject insulin through a needle placed under the skin. The first such commercial pump, Dean Kamen's AutoSyringe, was introduced in the late 1970s, but patients still had to program them based on regular blood glucose measurements taken through their fingers.
During this time, the patient continued to rely on fingertips. Finally, in 1999, Medtronic introduced the first portable continuous blood glucose monitor, sufficient for outpatient use. The thin electrodes are inserted under the skin with a needle, and then connected to a monitor worn on the body.
Abbott and Dexcom soon followed up with devices that display glucose data in real time. In the past 20 years, the accuracy of this instrument has been improving, and it is precisely because of these advances that artificial pancreas has become possible.
The ultimate goal is to replicate the entire work of the pancreas control system so that patients no longer need to manage themselves. But it turns out that imitating a healthy pancreas is extremely difficult.
Fundamentally speaking, blood sugar management is an optimization problem. Diet, exercise, disease, and other external factors that may affect metabolism make blood sugar management complicated. In 1979, the biomedical engineers Richard Bergman and Claudio Cobelli proposed the basis for solving this problem. They described the human metabolic system as a series of equations. However, in practice, it is difficult to find a solution due to three main reasons:
Delayed insulin action: In the body, insulin is secreted in the pancreas and shunted directly into the blood. But when injected subcutaneously, even the fastest insulin takes 40 minutes to an hour to reach its peak of action. Therefore, the artificial pancreas's controller must plan to lower blood sugar one hour from now-it must predict the future.
Inconsistent: The effect of insulin varies from person to person, even for the same person at different times.
Sensors are inaccurate: Even the best continuous blood glucose monitors can make mistakes, sometimes drifting in a certain direction-indicating that blood glucose levels are either too low or too high, and this problem may last for hours.
The artificial pancreas reproduces the glucose control system of a healthy body. The system begins when carbohydrates are digested into glucose and transported to the pancreas through the blood. The pancreas senses the increase in glucose concentration and secretes enough insulin to enable the body's cells to absorb glucose.
Two control systems based on the pancreas cooperate with each other to keep blood glucose levels within a healthy range. One uses insulin to lower high levels of glucose, and the other uses another hormone called glucagon to raise low levels. Today's artificial pancreas depends only on insulin, but two hormonal systems are being studied. Chris Philpot
More importantly, the system must take into account complex external influences, so it is as effective for a middle-aged man who sits at a desk all day and a teenager who sprints down the mountainside on a snowboard.
In order to overcome these problems, researchers have proposed various solutions. The first attempt was a direct proportional integral derivative (PID) controller, where the delivery of insulin is proportional to the increase in blood glucose level and its rate of change. After several improvements to the algorithm used to adjust the PID response to the speed of subcutaneous insulin delivery, the method is still used by a commercial system from Medtronic. A more complex method is the predictive control algorithm, which uses models of the human metabolic system, such as the model proposed by Bergman and Cobelli in 1979. The focus is to predict the future state, thereby partially compensating for the delayed diffusion of subcutaneous insulin into the bloodstream.
Another experimental controller uses two hormones-insulin, which lowers blood sugar levels, and glucagon, which raises blood sugar levels. In each of these methods, the modeling work creates a conceptual background for constructing an artificial pancreas. The next step is to actually build it.
To design a controller, you must have a way to test it. For this, biomedical engineering usually relies on animal testing. But such testing is time-consuming and expensive. In 2007, our team at the University of Virginia proposed to switch to computer simulation experiments.
Together with our colleagues from the University of Padua, Italy, we created a computer model of glucose-insulin dynamics that operated on 300 virtual subjects with type 1 diabetes. Our model describes the interaction of glucose and insulin over time through a differential equation that represents the best available estimate of human physiology. The parameters of the equation vary from subject to subject. The complete array of all physiologically feasible parameter sets describes the simulated population.
In January 2008, the U.S. Food and Drug Administration (FDA) made an unprecedented decision to accept our simulator as an alternative to animal testing in preclinical testing of artificial pancreas controllers. The agency agrees that this computer simulation is sufficient to obtain regulatory approval for inpatient human trials. Suddenly, fast and cost-effective algorithm development became possible. Only three months later, in April 2008, we started using the controller we designed and tested on a computer on real type 1 diabetic patients. The UVA/Padua simulator is now used by engineers all over the world, and animal experiments for testing new artificial pancreas algorithms have been abandoned.
Perhaps one day it makes sense to implant an artificial pancreas into the abdominal cavity, where insulin can enter the bloodstream directly to take effect faster.
At the same time, funding for research on other aspects of the artificial pancreas is also expanding. In 2006, JDRF (formerly known as the Juvenile Diabetes Research Foundation) began to develop a device in multiple centers in the United States and Europe; in 2008, the National Institutes of Health initiated a research program; from 2010 to 2014, the European Union The sponsored AP@Home Alliance is very active. The global boom in rapid prototyping and testing has yielded results: the first outpatient study was conducted in camps for children with diabetes in Israel, Germany, and Slovenia from September 2011 to January 2012. In these camps, children with type 1 diabetes were based on Artificial pancreas system for laptops.
Most of these early studies rated the artificial pancreas system as superior to manual insulin therapy in three respects. Patients spend more time within the blood sugar target range, have less hypoglycemia, and they have better control during sleep-when low blood sugar levels are difficult to detect and manage. But these early experiments all relied on laptop computers to run algorithms. The next challenge is to make the system mobile and wireless so that it can be tested under realistic conditions.
Our team at UVA developed the first mobile system, the Diabetes Assistant, in 2011. It runs on an Android smartphone, has a graphical interface, and is capable of web-based remote observation. First, we tested it in an outpatient setting in a study that lasted from several days to six months. Next, we tried it on high-risk patients because they often or severely experience hypoglycemia. Finally, we conducted a stress test on children with type 1 diabetes who were learning to ski in a 5-day camp.
In 2016, the key trial of MiniMed 670G, the first commercial hybrid system, ended. The system automatically controls the continuous rate of insulin throughout the day, but does not control the additional insulin doses injected before meals. The system was approved by the FDA for clinical use in 2017. Other groups around the world are also testing such systems, and the results are very good. A 2018 meta-analysis of 40 studies (1,027 participants) found that by comparison, patients stayed in the blood glucose target range (70-180 mg/dL) during sleep increased by approximately 15%, an overall increase of nearly 10%. Patients receiving standard treatment.
The third-generation offspring of our original machine — based on Control-IQ technology and manufactured by Tandem Diabetes Care in San Diego — conducted a six-month randomized trial in adolescents and adults with type 1 diabetes aged 14 years and older. We published the results in the New England Journal of Medicine in October 2019. The system uses Dexcom G6 continuous blood glucose monitor-no longer need to be calibrated by fingertip samples-insulin pump from Tandem, and the originally developed control algorithm UV. The algorithm is built into the pump, which means that the system does not require an external smartphone to process calculations.
Control-IQ software calculates additional insulin doses (called corrected boluses) to predict the increase in glucose concentration, reaching more than 162 mg per ten liters of blood. Corrections can be made every hour as needed. This is a supplement to the continuous infusion of insulin throughout the day, called the basal rate, which changes every 5 minutes according to the person's insulin needs.
The Minimed 770G artificial pancreas is a hybrid system that manages the metabolic insulin dose-it adjusts the basal rate, but does not manage the correction pills. It originated from the first such system approved for general use.
Control-IQ still requires some user involvement. Its hybrid control system requires people to press a button to say "I am eating" and then enter an estimated amount of carbohydrates; the person can also press a button to say "I am exercising." These interventions are not absolutely necessary, but they can make control better. Therefore, we can say that today's controllers can be used for full control, but they are better as hybrid vehicles.
The system has a dedicated safety module that will stop or slow down the flow of insulin whenever the system predicts hypoglycemia. In addition, it will gradually increase the insulin dose overnight to avoid the tendency of high points in the morning, and restore blood glucose levels to normal before 7 in the morning
The six-month trial tested Control-IQ against standard treatment. In standard treatment, patients used information from a blood glucose monitor to operate an insulin pump and complete all the work. Participants who used Control-IQ increased the time spent in the target blood glucose range by 11% and reduced the time spent below the low glucose red line (70 mg/dL) from 2.7% to 1.4%. In December 2019, the FDA authorized Control-IQ for clinical use in patients 14 years and older, so our system became the first "interoperable automatic insulin delivery controller" that can be connected to various insulin pumps and Continuous blood glucose monitor. Patients can now customize their artificial pancreas.
1. Two types that have been approved by the FDA:
Control-IQ from Tandem Diabetes Care
10. There are many, many ongoing DIY projects.
The day after experts in the Maryland conference room said the problem could not be resolved, the FDA approved it for nearly 14 years. One month after approval, Control-IQ was released as an online software upgrade to Tandem insulin pump users. In June 2020, following another successful clinical trial in children with type 1 diabetes between the ages of 6 and 13, the FDA approved Control-IQ for children 6 years and older. Compared with any other age group, children can benefit more from this technology because they are the least able to control their insulin dose.
In April 2021, we published an analysis of 9,400 people who used Control-IQ for a year. These real data confirmed the results of the early trials. As of September 1, 2021, Control-IQ has been used by more than 270,000 diabetic patients in 21 countries/regions. So far, these people have logged in on this system for more than 30 million days.
A parent wrote an article to Tandem about how Control-IQ greatly reduced his son's average blood glucose concentration in eight weeks. He wrote: "For this moment to come, I have waited and worked hard for 10 years." "Thank you."
Progress towards better automatic control will be gradual; when patients never intervene, we expect a smooth transition from hybrid to fully autonomous. Clinical trials are currently underway, and work on using fast-acting insulin is underway. Perhaps one day it makes sense to implant an artificial pancreas into the abdominal cavity, where insulin can enter the bloodstream directly to take effect faster.
What's next? So, what else is impossible today?
This article appears in the December 2021 print edition as "Creating an Artificial Pancreas".
Your weekly selection of wonderful robot videos
Video Friday is your weekly selection of wonderful robot videos, collected by IEEE Spectrum Robotics friends. We will also publish a weekly calendar of upcoming robotic events in the coming months; this is what we currently have (send us your events!):
If you have any suggestions for next week, please let us know and enjoy today’s video.
We first saw Cleo Robotics at CES 2017, when they were showing a consumer prototype of their unique ducted fan drone. They just announced a new version of enhanced surveillance, which is actually called Dronut.
For such a small matter, the 12-minute flight time is not the worst. I hope it can find a unique niche market and help Cleo return to the consumer market because I want one.
This is some very, very impressive robust behavior on ANYmal, which is part of Joonho Lee's master's thesis at ETH Zurich.
The title of this DeepRobotics video is "The End is Coming." It's better not to think about it, maybe.
At Ben Gurion University of the Negev, they are trying to figure out how to make the COVID-19 officer robot authoritative enough so that people can really pay attention to it and do what it says.
You would think that high-voltage wires are the last thing you want to let drones fly, but here we are.
This is probably the highest speed multiplier I have seen in the robot video.
This is an interesting manipulator design of Yale University Grablab, which can be easily operated by hand.
The ugo robot that is just a ball with eyes on a stick is one of my favorite robots because it is unapologetically just a ball on a stick.
Robot, make me a sandwich. Then make me a bunch of sandwiches.
Refilling water bottles is not a very complicated task, but letting robots do it means that humans don't have to do it.
The alphabet mapping tool shows atmospheric water harvesting in some of the driest areas in the world
Prachi Patel is a freelance journalist based in Pittsburgh. She writes articles on energy, biotechnology, materials science, nanotechnology and computing.
X, The Moonshot Factory, developed digital tools to assist so-called atmospheric water collection (AWH)-usually through solar water collectors and collectors, as shown in the picture.
One in four people in the world, about 2.2 billion people, cannot reliably obtain safe drinking water. According to a recent study published in the journal Nature, small solar devices that extract water from thin air can help 1 billion of them provide drinking water.
This new study by researchers at Alphabet, Inc.'s X, The Moonshot Factory (formerly Google X) maps the global potential of atmospheric water harvesting (AWH) technology for the first time.
"This is exciting because solar atmospheric water harvesting devices with unique potential must be off-grid, portable, low-cost, and renewable energy to put the power of drinking water sources in the hands of individuals," X said engineer Philipp Schmaelzle and His colleague Jackson Lord led the work together.
Some parts of the world are implementing desalination and wastewater treatment for the production of drinking water to produce drinking water on a large scale, but these technologies are still expensive for many others who lack financial resources and the required infrastructure.
Atmospheric water collectors that capture moisture from the air will be easier to use. But they have been largely ignored because, Schmaelzle said, “Historically, there has been some debate about whether the output of atmospheric water collectors is high enough to make them a viable solution for many people.”
Researchers estimate that AWH can theoretically provide enough water extraction yield to meet the basic needs of 1 billion people.
Various types of AWH systems have been designed by groups around the world. Passive devices usually rely on special fabrics and materials that collect dew or mist, and are limited to damp areas. At the same time, active devices use adsorbent materials to capture water vapor from the air, and use low-temperature solar heat to release water or cool to condense water vapor and produce water.
These solar-powered AWH systems can take advantage of natural sunlight and produce fresh water in areas with a relative humidity as low as 20%, thereby opening up a wider area on a global scale, where they can be used and have an impact.
To estimate the extent of this impact, Schmaelzle and his colleagues created a geospatial tool called AWH-Geo on the Google Earth Engine platform. The tool uses data from ERA5-Land, which is a publicly available data set that provides decades of high-resolution, hourly-evolving land variable information.
AWH-Geo looks at three variables to calculate the output of AWH equipment: solar irradiance, or solar energy illuminating a square meter of land; relative humidity; and air temperature. Researchers use published performance data of various AWH devices to plot their maximum theoretical output in different regions of the world.
Then, they compared the water production map with a WHO/UNICEF map of 2.2 billion people without access to safe drinking water. The amount of water produced by any AWH device in the real world may be less than its theoretical potential. But the researchers found that assuming more than 600 watts of solar radiation, assuming that a one square meter device can produce 0.2-2.5 liters/kWh of water, it can meet the average daily drinking water demand of about 1 billion people for 5 liters. Continuously supply power for two to three hours a day to collect water with a relative humidity of more than 30%.
Team X developed its own prototype AWH, which is suitable for cooling-condensing cycles. The device has a solar heating panel that can heat the air drawn by the fan. The water-absorbing material helps this hot air stream absorb moisture from the second air stream sucked in by another fan. The now moist, warm air stream is then sent to the lower chamber for cooling, causing water vapor to condense and collect at the bottom.
In the roof test, the equipment produced 150 milliliters of water per square meter of equipment area per hour at a cost of 10 cents per liter. This is enough to keep one person hydrated in many dry areas of the world. Further development may result in the production of 5 liters of equipment per day at a price as low as 1 cent per liter. But this will require more engineering and mass production, which the company says is beyond its vision. Instead, they make the design free for further development by others.
Startup SOURCE (formerly Zero Mass Water) has an impressive list of supporters, and similar devices are already on the market. However, for a two-panel array that produces about 7.5 liters of water per day, the price is still very expensive, about $6,000.
The X team’s analysis shows that the adsorbent-based technology uses large beds of advanced adsorbent materials to absorb moisture and solar energy to release steam, which may be more promising. High-grade adsorbent materials have the highest water output. But they are also expensive and require further development and mass production to achieve cost targets.
"Solar-driven atmospheric water harvesting is still an emerging technology," Schmaelzle said, "has not been adopted and implemented on a large scale. We hope that [our] the findings and tools of the paper can help inform the development of atmospheric water harvesting in the future Equipment so that researchers and designers can maximize the real-world impact."
November 19, 2021 correction: This story has been updated to emphasize the geolocation tool developed by Alphabet, rather than any specific atmospheric water trap device-previous versions of this story inaccurately emphasized this.
To understand whether your idea is new and patentable, you need to have the right tools to study the technological landscape. IP.com's artificial intelligence semantic search engine and suite of analysis and visualization tools allow you to easily understand how your ideas perform against the competition.