Design of a biosensor for continual, transdermal glucose monitoring.
To address these needs, a continual, noninvasive glucose monitoring system has been developed that utilizes reverse iontophoresis to induce an electroosmotic flux of glucose through intact skin. The electroosmotic glucose flux (measured ex vivo by HPLC) was shown in earlier clinical trials to correlate with the blood glucose concentrations (2). To create a miniaturized and wearable system, an amperometric biosensor has been developed that measures in situ the glucose extracted through the skin into a hydrogel pad to provide periodic measures of blood glucose every 20 min over the course of 12 h. A diagram of this system is shown in Fig. 1A.
A major advancement that has been achieved in the development of this biosensor is the ability to accurately measure the small amounts of glucose that are extracted through the skin. These glucose concentrations are several orders of magnitude lower than those present in the blood (~5 [micro]mol/L vs ~5 mmol/L), representing a total amount of glucose in the 50-200 picomol range. The main design criteria are summarized below.
Low concentration of glucose. The small amount of glucose extracted requires a biosensor with a detection limit much lower than conventional devices attain. A low detection limit has been achieved through the use of a large surface area working electrode and the development of a screen-printable platinum-graphite composite electrode material with high sensitivity and low background current. The large electrode, coupled with coulometric measurements, allows detection of a large fraction of the glucose collected into the hydrogel pad every cycle.
The characteristics of the biosensor were investigated in an in vitro cadaver skin diffusion cell thermostated at 32[degrees]C. This cell consists of a donor solution compartment separated from the biosensor/iontophoresis assembly by a layer of heat-separated human epidermis. This test mimics very closely the in vivo performance of both the iontophoresis and biosensor functions. In this test, the integrated biosensor current responded linearly when the glucose concentration in the donor solution was increased stepwise through the range from 0 to 5000 mg/L (0 to 500 mg/dL) glucose [mean correlation coefficient (r) over six separate systems, 0.992; range, 0.965-0.999].
Continual measurements. The continual, periodic measurement cycle requires that the sensing be completed and that the previously extracted glucose be depleted for the next iontophoresis cycle to begin anew. Each measurement cycle is made up of two half-cycles, each consisting of 3 min of iontophoresis and 7 min of biosensing. The polarity of the iontophoretic current alternates for each half-cycle. The sensor geometry, diffusion distances through the hydrogel pad, glucose oxidase enzyme kinetics, glucose mutarotation kinetics, and electron transfer kinetics on the electrode all affect the dynamics of the measurement cycle. These parameters have been co-optimized using computer modeling (3) to achieve high sensitivity glucose measurement and subsequent depletion of remaining glucose within the required time period.
The sensitivity of the system must be stable to make repeated measurements over 12 h. Stability was investigated in nine separate in vitro diffusion cell systems. Each system was calibrated at the beginning of the measurement period to account for skin flux differences and the usual variances in the sensitivity of the biosensor. Three systems each were stepped to glucose concentrations of 500, 1000, and 4000 mg/L (50, 100, and 400 mg/dL), and the stability was evaluated over 36 measurements over the course of 12 h (Fig. 1B). The stability, as measured by the mean CV for the three cells at each concentration was 7.7%, 5.2%, and 5.1% for 500, 1000, and 4000 mg/L concentrations, respectively.
[FIGURE 1 OMITTED]
Selectivity to glucose. The biosensor has low sensitivity to most potential interfering species. This selectivity was achieved by utilizing the inherent permselective properties of the skin and the electroosmotic extraction process. The extraction through skin provides both size- and charge-exclusion properties. The size-exclusion properties arise from the molecular weight cutoff of ~500 for efficient extraction of compounds through the skin (4). The charge-exclusion properties are caused by the net negative charge carried by skin at physiological pH (5). Because of this charge, the electroosmotic flux, which carries the glucose, flows predominantly toward the iontophoretic cathode. Anionic species, such as the interfering species ascorbate and urate, migrate solely to the iontophoretic anode. By sensing glucose only at the biosensor at the iontophoretic cathode, the biosensor avoids interference from these species. Finally, by integrating the biosensor signal over the measurement cycle, the biosensor can operate at a fairly low potential (0.42 V vs a Ag/AgCI electrode). This potential is much lower than is typically used in glucose sensing and is at a value where some interfering species no longer react at the electrode. For example, background currents from tyrosine and tryptophan, which are seen at 0.6 V, are completely eliminated at 0.42 V.
Toxicity. Because the device is worn on the skin, the iontophoretic process would effectively deliver any mobile chemical species into the skin. This requirement precludes the use of sensing chemistries containing soluble mediators and other schemes incorporating mobile toxic or irritating compounds. The sensing chemistry and all other components of the device must pass strict skin toxicity and irritation tests.
The G1ucoWatch[R] biographer (developed by Cygnus, Inc., Redwood City, CA) embodies the electroosmotic extraction and biosensor system described above in a small, wristwatch device. This device has been tested extensively in clinical trials on diabetic subjects and has been shown to accurately measure blood glucose in a continual and noninvasive manner.
(1.) The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus, N Engl J Med 1993;329:997-1036.
(2.) Tamada JA, Bohannon JV, Potts R0. Measurement of glucose in diabetic subjects using noninvasive transdermal extraction. Nat Med 1995;11: 1198-201.
(3.) Kurnik RT, Berner B, Tamada J, Potts R0. Design and simulation of a reverse iontophoretic glucose monitoring device. J Electrochem Soc 1998;45: 4119-25.
(4.) Green PG, Hinz RS, Kim A, Szoka FD, Guy RH. Iontophoretic delivery of a series of tripeptides across the skin in vitro. Pharm Res 1991;8:1121-7.
(5.) Luzardo-Alvarez A, Rodriguez-Fernandez M, Blanco-Mendez J, Guy RH, Delgado-Charro MB. Iontophoretic permselectivity of mammalian skin: characterization of hairless mouse and porcine membrane models. Pharm Res 1998;15:984-7.
Michael J. Tierney, * Yalia Jayalakshmi, Norman A. Parris, Michael P. Reidy, Christopher Uhegbu, and Prema Vijayakumar (Cygnus, Inc., Redwood City, CA 94063; * author for correspondence: fax 650-369-5318, e-mail Tierney@cygn.com)
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Abstract of Oak Ridge Posters|
|Author:||Tierney, J. Michael; Jayalakshmi, Yalia; Parris, A. Norman; Reidy, Michael P.; Uhegbu, Christopher;|
|Date:||Sep 1, 1999|
|Previous Article:||Management of interferences in a transdermal, noninvasive glucose monitoring device.|
|Next Article:||Development of kinetic ligand-binding assays using a fiber optic sensor.|