Toward Continuous Blood Glucose Monitoring

Originally Published MDDI June 2003

Glucose Monitoring

For more-effective diabetes management, glucose monitoring should be minimally invasive, continuous, unobtrusive, and informative. Meeting this goal requires microminiature sensors and electronics. MEMS fabrication, new materials, and IC assembly techniques are now making it a practical reality.

by Albert P. Kretz and Donald Styblo

For years, diabetics have monitored their glucose level by drawing blood using the finger-prick method. Because of the pain and inconvenience of manipulating several diagnostic kit items, most do not comply as often as they should. Type I diabetics usually test only one or two times per day, instead of the four to six times required for optimal glucose control. This often results in uneven insulin usage, sometimes with life-threatening results. Newer diagnostic techniques requiring less blood now allow sampling at less-sensitive sites on the arm or leg, and do so with a small volume of fluid. While this encourages more frequent monitoring, it produces only a snapshot, and is nowhere close to continuous.

Failure to manage diabetes properly has dire consequences for individuals and for society as a whole. Estimates by the World Health Organization and other sources peg the worldwide diabetic population at about 176 million people. It is expected to grow to 370 million by 2030. This increase is much higher than the general population growth, particularly in the United States. Key factors in this trend are poor diet, lack of exercise, increasing obesity, and population aging. Many of those with Type II diabetes, the most prevalent type, do not even know they have the disease. Untreated, the disease leads to serious microvascular consequences that include blindness, renal failure, coronary artery disease, and limb amputations.

According to the results of the 1993 Diabetes Control and Complication Trial, tight glucose monitoring can help reduce these complications by 40 to 75%. In addition, a convenient continuous monitoring system would be valuable in diagnosing potential cases of Type II diabetes. The direct and indirect costs of diabetes exceed $132 billion annually in the United States—about 20% of all healthcare costs. Improving diabetes management can provide enormous health and economic benefits. Federal and state governments recognize this, and many states mandate that health insurance providers pay for diabetes diagnostic instruments and supplies.

Barriers to Better Monitoring

A painless, noninvasive or minimally invasive technique is the holy grail of many companies that produce, or aspire to produce, blood glucose monitoring products. Their efforts thus far have produced mixed results, particularly where continuous monitoring is concerned.

Noninvasive Methods. Most noninvasive methods use some type of infrared (IR) light sensor to detect blood glucose levels without puncturing the skin. These devices typically shine light on the skin and then measure the IR absorption or transmission wavelength specific to glucose. Calibration of these devices has proven to be a problem because of other chemicals in the blood that have properties similar to glucose and interfere with the reading. For some designs, the calibration period can be quite long (up to 60 days). It also must be specific to the individual to achieve a useful signal-to-noise ratio. Such design and calibration requirements make for high unit cost, so market success may not be likely. 

Other optical methods use various wavelengths, measurement sites, and electronic circuitry. A few devices based on these techniques have undergone limited human trials and produced reasonably accurate results compared with plasma blood glucose measurements. Few if any of these optical designs have received FDA endorsement for an expanded clinical trial.

Likewise, electromagnetic waves have been used to sense blood glucose through the skin noninvasively via radio-wave impedance spectroscopy. By varying radio-wave frequencies, changes in blood glucose can be determined. Again, calibration can be problematic, but at least one of these devices is at the early trial stage.

There also are sampling techniques that are essentially noninvasive, since they use interstitial fluid (ISF) drawn through virtually intact skin. One sampling method uses reverse iontophoresis (electro-osmosis) to extract the ISF. Another uses a laser to painlessly create micropores in the dead layer of skin, through which ISF is collected. The latter technique may require an additional mechanism to draw the fluid, because ISF does not always flow freely into the pores. In any case, glucose in the ISF can be correlated with blood glucose. Initial calibration requires blood glucose measurements obtained with the finger-stick method. 

To complete an ISF glucose measurement, the general approach is to have the glucose molecules in the fluid react in a gel-collection disk. This produces an electrical signal proportional to the blood glucose level. The disk usually is attached to the skin as an adhesive patch. Perspiration, lag time, and other factors can affect measurement accuracy. An advantage of this method is the small electronic module (about the size of a wristwatch) for reading and storing the electronic data. In one of these devices, up to three measurements per hour for 12 hours are possible. Then the gel disk must be replaced. While this technique is not fully continuous, it is an improvement over finger-stick sampling. FDA has approved it for patients aged 18 and older.

Minimally Invasive Methods. A truly minimally invasive technique can be rigorously defined as one that does not use subcutaneous sampling or sensors in fatty tissue to collect blood. Rather, percutaneous needles or sensors enter the dermis to collect or react with an ISF or blood sample, and measure glucose as described above. While the dermal layer has many capillaries, it has fewer nerve endings at sites other than the fingers. So, if the sampling needle or sensor diameter is small enough, there is no pain in the usual sense.

To make continuous glucose monitoring practical, the sampling device or sensor must remain in place for at least a few days. Proposed methods of doing this include tiny sensors or sampling probes pushed into the dermis and similar devices inserted just under the skin. Typical measurement technologies use an electrochemical sensor based on a glucose oxidase (GOX) or similar reaction. With suitable wireless radio technology, readings can be sent to a user's data storage and display device for an immediate update.

There are no truly minimally invasive devices like the ones described above approved by FDA for sale in the United States. Before this can happen, there are a number of technical problems to overcome. In addition to basic calibration and accuracy issues, the design of a continuous glucose monitoring system (CGMS) must overcome the barriers described in Table I.

Tearing Down Barriers

Recently, a biotechnology consortium began work to remove the barriers to a practical CGMS. The specific goal is a cost-effective system that is minimally invasive and convenient, and that provides real-time results accurately calibrated to blood glucose levels. The consortium is led by Advanced BioSensors Inc. (ABSI) of Cleveland. It includes experts in diabetes, diagnostic products, biomedical sensors, enzymatic reactions, biocompatible materials, MEMS (microelectromechanical systems), and electronics microminiaturization. 

The miniature scale of physical devices and processes is the principal challenge (and key to success) in combining these technologies in a practical CGMS. Reliable measurements can be achieved with a GOX-based electrochemical sensor. But to make such a sensor minimally invasive and virtually painless, it must be only 1 to 2 mm long by 0.1 to 0.2 mm wide. It requires microfabrication techniques to create a device that can achieve a high level of enzyme loading and measurement sensitivity in the small space available. By combining MEMS fabrication methods with certain polymer technologies, it is possible to bond GOX to a small sensor surface and produce an acceptable output signal.

To counter adverse body reactions, ABSI has developed a design that will use a proprietary biochemical coating on the sensor. This will allow the sensor to remain in place for 3 to 7 days. The biochemical coating has sufficient permeability to allow an appropriate level of GOX reaction with the blood glucose in the sample area. Within the planned implant time span, any deterioration in sensor output can be compensated for electronically.
The sensor output is an electrical current that can be correlated with the glucose concentration. Benchtop prototypes have been produced and successfully tested, as shown in Figure 1. The time required to get a stable output is less than 1 minute for virtually all clinically significant glucose concentrations (Figure 2). A provisional patent application has been filed on the sensor platform technology, which covers additional diagnostic and therapeutic applications.

The sensor may be mounted on a disposable adhesive patch. This allows its use anywhere on the body and minimizes skin irritation from continuous use at a single site. The needle-shaped sensor on the patch is merely patted into place in a virtually painless application. The patch will have a docking port for an electronics assembly that consists of a low-noise, high-input impedance amplifier, analog-to-digital convertor (ADC), wireless transceiver, and power source (Figure 3).

The encrypted digital data will be transmitted to a wristwatch-sized monitor/recorder module. The monitor electronics include a power source and wireless transceiver, microprocessor, programmable RAM for data storage, and an LCD panel to display the blood glucose reading. The display will be kept simple, so that even children and the elderly can easily understand the results. Alarms will be provided for glucose levels that are too high or too low. To help avoid a crisis, an alarm will also sound if the patient's glucose level is trending up or down too fast. It is anticipated that the monitor transceiver will have an encrypted output signal level large enough to allow secondary wireless connectivity to a remote physician's office.
This system will allow virtually continuous monitoring, limited only by the settling time required for each measurement (less than a minute). Measurement frequency may be adjusted based on the size of the data storage memory employed. A first-in/first-out data deletion scheme can be used to retain the latest readings in memory. The data can be recalled as required for local display or transmission to a healthcare provider.

In production, most of the electronics will be implemented in application-specific integrated circuits (ASICs). Valtronic USA Inc. (Solon, OH) will use its flip-chip assembly process to develop the microminiature modules that incorporate the ASICs, power sources, and related circuit components that make up the sensor patch and monitor/recorder electronics. 

Trials on the Horizon

Currently, a prototype MEMS production process is being developed that will convert base sensors into chemical-sensitive units on a wafer produced with methods similar to integrated circuit fabrication. One goal of this production technique is a sensor price comparable to consumable supplies associated with finger-stick glucose monitoring.

In the next step, Valtronic will separate the sensors from the wafer, mount each one with electrical connections, and add the biocompatible polymer before assembling the sensor patch. In addition, the microminiature circuitry for encrypted wireless connectivity between the patch and monitor/recorder will be assembled. Sterilization techniques will be developed for this package. At that point, a small animal trial will begin.
After completion of second-round financing, the plan is to finalize the product design, including wireless connectivity to the monitor/recorder. Additional financing will support clinical trials and a regulatory (PMA) submission. Looking further into the future, the aim is to combine the CGMS with an insulin pump to produce a closed-loop system—an artificial pancreas.

Acknowledgments

This project was supported by Advanced BioSensors Inc., with generous cooperation and test data from Alain Izad, PhD, AMMI division of B.F. Goodrich Corp.; and Miklos Gratzl, PhD, Koji Tohda, PhD, and Jian Yang, PhD, of the Biosensor Laboratory, Department of Biomedical Engineering, Case Western Reserve University, Cleveland.

 Copyright ©2003 Medical Device & Diagnostic Industry