Ingestible Capsule for Impedance and pH Monitoring in ... - IEEE Xplore

IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 54, NO. 12, DECEMBER 2007

2231

Ingestible Capsule for Impedance and pH Monitoring in the Esophagus Jose L. Gonzalez-Guillaumin, Daniel C. Sadowski, Karan V. I. S. Kaler, Member, IEEE, and Martin P. Mintchev*, Senior Member, IEEE

Abstract—Twenty-four-hour ambulatory pH monitoring is an essential tool for diagnosing gastroesophageal reflux disease (GERD). Simultaneous impedance and pH monitoring of the esophagus improves the detection and characterization of GERD. Conventional catheter-based monitoring systems are uncomfortable and interfere with the normal activity of the patient. To overcome these disadvantages, different wireless esophageal monitoring systems have been proposed. A capsule containing sensors for impedance and pH monitoring with wireless communication capabilities is presented. A low cost miniature microcontroller was utilized for interfacing between the sensors and a wireless transmitter. The microcontroller program allowed efficient management of the electric power provided by a 3-V battery. Magnetic holding is proposed as an alternative to surgical affixation of the monitoring capsule. Permanent neodymium magnets separated by 27 cm successfully held the capsule in a test tube. Experimental results demonstrated that friction force can aid magnetic holding to overcome peristalsis. The proposed design efficiently detected acid and nonacid reflux. More research regarding the holding method and capsule packaging are necessary to optimize the mechanical performance of the proposed design in order to facilitate clinical testing on human subjects. Index Terms—Esophageal sensor, gastro-esophageal reflux, impedance sensors, pH sensor.

I. INTRODUCTION ASTROESOPHAGEAL reflux disease (GERD) is characterized by the reflux of gastric content back into the esophagus. Traditionally, pH monitoring alone has been used to diagnose GERD. Catheters for reflux monitoring use a single pH electrode located 5 cm above the lower esophageal sphincter (LES) and can effectively detect acidic reflux [1]. However, nonacidic reflux cannot be detected using a pH electrode alone. Recent studies have demonstrated that reflux symptoms can also occur in association with nonacid reflux [2]. Multichannel intraluminal impedance (MII) detects changes in conductivity

G

Manuscript received January 27, 2006; revised February 27, 2007. This work was supported in part by Sandhill Scientific Inc., Denver, Co., by The Gastrointestinal Motility Laboratory, Edmonton, AB, Canada, and by the Natural Sciences and Engineering Research Council of Canada. Asterisk indicates corresponding author. J. L. Gonzalez-Guillaumin and K. V. I. S. Kaler are with the Department of Electrical and Computer Engineering, University of Calgary, AB T2N1N4, Canada. D. C. Sadowski is with the Faculty of Medicine, University of Alberta, Edmonton, AB T6G2E1, Canada. *M. P. Mintchev is with the Department of Electrical and Computer Engineering, University of Calgary 2500 University Drive N.W., AB T2N 1N4, Canada, and also with the Faculty of Medicine, University of Alberta, Edmonton, AB T6G2E1, Canada (e-mail: [email protected]). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TBME.2007.908332

provoked by nonacidic reflux, acidic reflux, and bolus in the esophagus. Therefore, MII combined with pH monitoring provides valuable information to discriminate between acidic and nonacidic reflux [1]. Impedance-pH catheters are inserted transnasally into the esophagus. Thus, the main disadvantage of these catheters is the discomfort suffered by the patient. A University of Kentucky review of 520 ambulatory pH procedures showed that 4% of patients were unable to tolerate the intranasal catheter and did not complete the study [3]. Moreover, 10% of the patients to whom the test was recommended refused to take it. In addition, despite specific instructions to maintain a normal daily routine during the test period, most patients altered their routine and changed their eating habits due to the discomfort caused by the catheter [3]. Therefore, the quality and accuracy of these studies are potentially jeopardized. A wireless alternative to catheter-based pH and impedance monitoring for GERD, therefore, would be a valuable diagnostic tool. Different pH monitoring devices relying on wireless communications have been proposed to increase the patient comfort level. In addition to pH sensing capabilities, the “SmartPill Capsule,” offers sensors to monitor temperature and transit time [4]. However, this device is not optimized for 24-h ambulatory monitoring because it is not affixable to the esophageal mucosa above the LES. Usually, the pH sensor has to be affixed at a 5-cm distance from the LES. The “Bravo Capsule” by Medtronic is affixable to the mucosal wall of the esophagus using a needle. However, the needle is positioned and removed using an invasive endoscopic clinical procedure [5]. In addition, the Bravo Capsule cannot discriminate between acidic and nonacidic reflux. The Bravo Capsule contains an antimony pH electrode and transmits two pH data values every 12 s [5]. A remote-query magneto-elastic pH sensor has also been proposed [3]. However, it is unclear how this sensor can be affixed to the esophageal wall. This magneto-elastic pH sensor cannot discriminate between acidic and nonacidic reflux as well. The goal of this research was, therefore, to develop an innovative impedance-pH wireless capsule capable of discriminating between acidic and nonacidic reflux, which can be affixed on the inner side of the esophageal wall above the LES and with minimal discomfort to the patient. II. METHODS A. System Overview The proposed design integrates sensors for pH and impedance monitoring into an ingestible capsule. Stainless-steel half-ring electrodes were utilized to measure impedance and discriminate between acidic and nonacidic esophageal reflux. A standard

0018-9294/$25.00 © 2007 IEEE

2232


B. pH Sensing An ingot of metallic antimony, consisting of 99.999% metal basis (Alfa Aesar, Ward Hill, MA) was pulverized, melted, and chemically treated to form antimony billets of 1 mm in diameter and 2 mm in length. Silver wire (California Fine Wire, Grover Beach, CA) was chloridized, rinsed and baked at 175 C to be utilized as a reference electrode. An antimony billet was soldered to a copper wire connected to a circuit board. The reference electrode was coated in gel (Signa Gel, Parker Laboratories, Inc., Fairfield, NJ) and encapsulated next to the antimony billet [Fig. 1(b)]. The fabricated pH sensor, together with the associated integrated electronics had dimensions of 2.2 mm in diameter by 20 mm in length. The difference in voltage between the reference electrode and the antimony billet depended on the pH value. A miniature low power operational amplifier (LMV981BL, National Semiconductors Corporation, Santa Clara, CA) was used as a buffer between the pH sensing electrodes and the microcontroller A/D converter. This miniature operational amplifier requires a footprint of 1.006 mm 1.514 mm 0.954 mm. The reference electrode was connected to the ground through a 14-Quad-Flatpack-No-Lead (14 QFN) analog switch (SN74LV4066ARGYR, Texas Instruments Inc., Dallas TX) to minimize interference provoked by the impedance stimulating pulse. C. Impedance Sensing

Fig. 1. (a) Impedance-pH capsule and (b) its Antimony billet.

antimony electrode containing an internal reference electrode and conductive gel was utilized to sense and monitor pH values [1]. A miniature microcontroller was employed to perform analog-to-digital (A/D) conversion of the data, power management, and wireless communication control (Attiny26, Atmel Corporation, San Jose, CA). A radio frequency (RF) transmitter allowed wireless communication between the capsule and a computer station (MAX1472, Maxim Integrated Products Inc., Sunnyvale, CA). One 9-mm, 30-mAh 3-V battery supplied power to the system (CR927, Tian Qiu Corporation, China). A 6.5-mm 26-mm 1.5-mm neodymium magnet was included in the capsule to facilitate magnetic holding. Friction-enhancing pins were used to increase the static friction coefficient between the capsule and the mucosal wall of the esophagus to aid magnetic holding. All previously described components were integrated into a 28-mm 8-mm 8-mm capsule design with a weight of 5.9 g [Fig. 1(a)]. The capsule 10-cm was magnetically held in position using two 10-cm 5-cm, 5000 Gauss, square magnets positioned 27 cm apart, and the friction force between the pin-shelled side of the capsule and an acrylic model of the esophageal wall. The acrylic model of the esophageal wall was molded polyvinyl chloride tube simulating the esophagus of an average human, and had uneven inner lining mimicking the mucosa. The mucosal-like lining was formed during casting using a specially designed polystyrene foam core with uneven surface, over which the polyvinyl chloride tube dried.

Stainless steel half-ring electrodes were utilized to monitor impedance changes. Five cycles of a 100-Hz 3-V square pulse were generated utilizing the Attiny26 microcontroller. A 100-k resistor was used to limit the current flow through each of the electrodes. The sensing electrodes and the reference were connected to the microcontroller and ground through an analog switch. Two impedance channels were implemented using three 2.5-mm wide half-ring electrodes. The distance between the electrodes was set to 2.5 mm. The impedance-sensing capability was tested using a matrix of 10 standard 5% resistors ranging from 1 to 100 k . During this testing, the resistors were attached to the electrode rings. D. Capsule Circuitry Voltage changes in the pH and impedance sensors were monitored by the A/D converters of the Attiny26. The A/D converter noise reduction mode was utilized. This mode not only reduces noise but also reduces power consumption due to the fact that the digital blocks of the Attiny26 are turned off during the conversion time. The digital data from the A/D converter were stored in memory. Every 3 s these values were transmitted using the MAX1472. A block diagram of the circuitry used in the capsule is shown on Fig. 2(a). Fig. 2(b) depicts the fabricated 4-layer printed circuit board, which was subsequently housed in the capsule. E. Control Software The embedded software of the microcontroller was designed to configure the ports of the Attiny26. In addition, security identifications were added to each reading. To assure the integrity of the data during the wireless communication, a checksum test

GONZALEZ-GUILLAUMIN et al.: INGESTIBLE CAPSULE FOR IMPEDANCE AND PH MONITORING

2233

Fig. 3. Forces acting on the capsule.

Fig. 2. (a) Block-diagram and (b) printed circuit board of the proposed design.

was implemented. The microcontroller also determined the delays between transmissions in order to save power and generated all the necessary control signals. The delays between transmissions and security identifications can be easily programmed before assembling the capsule. Thus, these parameters can be optimized according to the needs of different clinical studies. Frequency encoding was implemented for the wireless communication. The Atmel in-system programmer (ISP) and Atmel Studio 4 software were used for programming the microcontroller (Atmel Corporation, San Jose, CA). F. Receiver Block Outside the body, a MAX1473 receiver kit was used (Maxim Integrated Products Inc., Sunnyvale, CA). The output of the receiver was also filtered and decoded by an Attiny26 microcontroller. The digital results were acquired into a computer and converted to pH and impedance values for real-time monitoring, data analysis and system debugging. A DAQCard AI-16XE-50 and C for Virtual Instrumentation (CVI) v. 7.0.0 were used to implement this stage (National Instruments Corporation, Austin, TX). G. Friction-Assisted Magnetic Holding Magnetically held devices for gastrointestinal (GI) monitoring have been proposed and tested in the past [6], [7]. Experimentally, it was determined that a field of 200 Gauss was sufficient to hold a 6-g mass in position against gravity. The human esophagus is located approximately at the center of the rib cage, behind the lungs and in front of the spinal cord and the aorta. The magnetic field generated by two 10-cm 10-cm

5-cm, 5000 Gauss magnets separated by a 27-cm gap was simulated in Maxwell 3-D CAD software (Ansoft Corporation, Pittsburgh, PA). Permanent neodymium magnets with the described characteristics were designed and fabricated (Magnet City, Miami, FL). Each magnet alone weighed 3.78 kg. The magnetic force holding the capsule in place can be combined with static friction between the shell of the capsule and the inner side of the esophageal wall to overcome the propelling . The static friction is defined as the reperistaltic force quired force to start moving a body at rest [8]. The static friction between two solid surfaces can be defined as the coefficient ratio of the tangential force required to cause the movement, [8] divided by the normal force between the surfaces

(1) The forces acting on the capsule are shown on Fig. 3. The magnitude and the angle of depend on how the peristaltic is force is distributed on the surface area of the capsule. because the dome-like shape of the capsule top is shown at experienced expected to cause this effect. The normal force by the capsule is the resultant of the “ ” component of the mag), plus the “ ” component netic force (i.e., (i.e., ). In the worst case scenario of (when the capsule is not supported by the esophageal wall), is expected to the “ ” component of the magnetic force overcome both gravity and the peristaltic force. The “ ” com) corresponds to ponent of FPP (i.e., in (1). Therefore, the maximal required static friction coefficient to hold the capsule against peristalsis can be determined as

(2) Manometric recordings obtained from healthy volunteers during peristalsis are shown in Table I [9]. The strongest provoked by peristalsis reported in this study pressure

2234


TABLE I TYPICAL PERISTALTIC PRESSURE VALUES IN 24-H AMBULATORY MONITORING OF ESOPHAGUS

was 96 mm Hg during meals. This corresponds to 12.8 kPa or 12800 N/m . The force propelling the capsule depends on the shape of the capsule and the angle between the surface of the capsule and the contracting esophageal wall at the point of contact. Therefore, a capsule housing design meeting propulsive proximal contractions at its top at an angle considerably reduces the propelling force acting on its surface. Peristaltic forces acting at a 90 angle to the top surface area of the capsule would not propel the capsule, but instead would only push it against the esophageal wall. This, however, would be impossible to achieve, taking into account the fact that the peristaltic wave propagates distally and would inevitably impact the proximal surface area of the capsule with a force at an angle smaller than 90 . If we consider the realistic assumption that this area is approximately 1/4 of the capsule can be estimated as total surface area, and at a 45 angle,

N

(3)

A force of 0.863 N corresponds to a mass of 88 g against gravity. The negative sign indicates the direction of the force. Therefore, a combination of the external magnetic field and friction-enhancing pins capable of handling a load of about 100 g without penetrating the mucosa should be able to overcome , and the feasibility of the approach could be tested in laboratory conditions simply by attaching a weight equivalent to to the capsule in a model of the esophagus. An array of 18 nonsharpened stainless steel pins (0.16 mm in diameter and 0.7 mm in length) was build by silver-soldering the pins to a stainless steel plate (Stay-Brite Silver Solder Kit, J. W. Harris Corporation, Manson, OH). The array was attached to the capsule and loaded with a 100-g mass made from moldable plasticin in order to test this friction-assisted magnetic holding in a laboratory setup. Particular care was taken so that the plasticin was not touching the walls of the laboratory model so that the friction was solely between the latter and the capsule. It is worth noting also, that a superposition between the peristaltic force and gravity in the esophagus would be an overestimation of the actual force acting on the capsule, since in a normal esophagus the walls of the organ are collapsed on each other, thus significantly marginalizing the impact of gravity. In a pathologically expanded distal mega-esophagus, the gravity alone could be considered a factor, but then the propulsive action of the peristaltic waves would be significantly diminished. Thus, a 100-g

Fig. 4. (a) Assembled capsule and (b) pH and impedance values transmitted by it.

force model applied to the capsule seems a reasonable overestimation of the actual displacing force that the capsule can experience in the esophagus. H. Experimental Setup The printed circuit board, the sensors and the magnets were assembled according to the proposed design [Fig. 4(a)]. The prototyped capsule was magnetically held in an polyvinyl chloride tube of 2.8 cm in diameter with specifically designed braided inner surface, simulating the geometry of the mucosal lining of a human esophagus. The designed capsule was sequentially submerged in pH buffers of 7 and 4. The sensor values were read, transmitted through an RF wireless link,

GONZALEZ-GUILLAUMIN et al.: INGESTIBLE CAPSULE FOR IMPEDANCE AND PH MONITORING

2235

and saved into a computer. Impedance measurements were validated using 10- and 20-k resistors. III. RESULTS A. pH Sensing Antimony electrodes require pre-calibration and drift compensation. Thus, a repeatability test could not be applied to the electrodes prior to proper calibration. However, a voltage change of 100 mV was generally observed between pH 7 and pH 4. After calibration, the designed pH sensor successfully detected pH changes in both pH buffers 4 and 7. The voltage value for pH of 7 was around 350 mV 20 mV and 250 mV 20 mV for pH of 4. A linear response was observed with a corresponding voltage of 33.3 mV per pH unit. The necessary calibration time was about 10 min. B. Impedance Impedance changes were successfully detected and corresponded to the industrial standard at the given pH values [2]. In addition, the readout impedances reflected the utilized test resistor values after calibration. The measurements were . repeatable C. Acidic and Nonacidic Reflux Fig. 4(b) depicts samples obtained during 330 s of experimental laboratory testing. The first 60 s corresponded to system synchronization. Therefore, during this time interval the readout impedance and pH values were meaningless. In the interval 60–170 s we observed a pH value of 4 and a corresponding impedance drop. In the subsequent 170–200 s interval the capsule was transferred from pH buffer 4 to 7. Thus, a rise in impedance and pH values was noted, bringing the signal close to saturation. From 210 to 330 s, a pH of 7 was monitored, and the corresponding impedance dropped. This last segment of the graph can be interpreted as a nonacidic reflux episode in which an impedance drop was registered, while the pH had a nonacidic value. D. Magnetic Holding Fig. 5(a) presents simulation results for the two separated 5000 Gauss magnets. The simulated field corresponded to the values measured experimentally. The capsule was successfully held in position during all experiments [Fig. 5(b)]. As expected, FM was strong enough to cancel FG. In the framework of a modified laboratory setup using polystyrene foam rather than the tubular esophageal model due to the lack of space to position the 100-g load, the friction-enhancing pins were able to hold the load against gravity [Fig. 5(c)]. Therefore, the proposed capsule design can theoretically overcome esophageal peristalsis while remaining affixed to the esophageal wall without penetrating the mucosal lining. IV. DISCUSSION A capsule design for combined impedance-pH monitoring has been presented and tested in laboratory conditions. In contrast to previously proposed solutions, this design was able to discriminate between acidic and nonacidic reflux. After the

Fig. 5. (a) Maxwell 3-D model of the forces surrounding the capsule, (b) its holding in the tubular experimental model, and (c) at polystyrene-supported 13.5-cm distance.

necessary synchronization and calibration, the obtained pH and impedance readings were stable and consistent. The utilized permanent magnets successfully held the capsule in position in an acrylic tube modeling the esophageal wall and anatomical distances between the permanent magnets and the capsule. However, the magnets utilized in the present experimental setup were bulky and heavy. Maxwell 3-D simulations using magnets of 20 cm 20 cm 1 cm, which delivered the same magnetic flux, offered similar results and behavior. These magnets would be 20% lighter, thus facilitating the design of a vest for human testing (Fig. 6). The vests should serve as an appropriate shield for the strong magnetic field as well. However, specific magnet design and implementation within the shielding vest would be required. Improvements to the capsule housing can also be

2236


[6] T. L. Abell and J. R. Malagelada, “Glucagon-evoked gastric dysrhythmias in humans shown by an improved electrogastrographic technique,” Gastroenterology, vol. 88, pp. 1932–1940, 1985. [7] Development of capsule endoscopes and peripheral technologies for further expansion and progress in endoscope applications, Olympus Corp., Tokyo, Japan, 2004 [Online]. Available: http://www.olympus.co.jp/en/news/2004b/nr041130capsle.cfm [8] H. D. Young and R. A. Freedman, University Physics, 9th ed. Menlo Park, CA: Addison-Wesley, 1996, pp. 132–134. [9] H. J. Stein, S. Singh, and T. R. Demeester, “‘Efficacy’ of esophageal peristalsis: A manometric parameter to quantify esophageal body dysfunction,” Dis Esophagus, vol. 17, pp. 297–303, 2004. Fig. 6. Proposed vest for ambulatory monitoring.

implemented in order to reduce the impact of peristalsis. In addition, plastic friction-enhancing pins would be preferable to the present stainless-steel solution. The proposed design was tested in stationary laboratory conditions. Clinical studies in dynamic conditions are needed to further assess the feasibility of the proposed approach. Should these studies prove successful, a mechanism to initially position the capsule 5 cm above the LES, preferably using a standard catheter carrier should be developed. V. CONCLUSION An innovative multisensor esophageal capsule design has been presented and tested in stationary laboratory conditions for the purpose of simultaneous detection of acidic and nonacidic gastro-esophageal reflux. The obtained results suggest that if appropriate shielding of the external permanent magnets is provided, and the latter are of appropriate weight and size, this technique could offer a minimally invasive and reliable testing of all aspects of GERD. A capsule suitable for clinical testing needs to be developed so that the proposed method is dynamically tested in real-life conditions. REFERENCES [1] P. J. Kahrilas and E. M. Quigley, “Clinical esophageal pH recording: A technical review for practice guideline development,” Gastroenterology, vol. 110, pp. 1982–1996, 1996. [2] A. J. Bredenoord, B. L. Weusten, R. Timmer, and A. J. Smout, “Reproducibility of multichannel intraluminal electrical impedance monitoring of gastroesophageal reflux,” Amer. J. Gastroenterol., vol. 100, pp. 265–269, 2005. [3] C. Ruan, K. G. Ong, C. Mungle, M. Paulose, N. J. Nickl, and C. A. Grimes, “A wireless pH sensor based on the use of salt-independent micro-scale polymer spheres,” Sensors Actuators B, vol. 96, pp. 61–69, 2003. [4] G. Schaefer, “Identification Pill With Integrated Microchip: SmartPill, SmartPill With Integrated Microchip and Microprocessor for Medical Analysis and a SmartPill, SmartBox, SmartPlate for Luggage Control on Comercial Airlines,” U.S. 5 792 048, Aug. 11, 1998. [5] J. E. Pandolfino, J. E. Richter, T. Ours, J. M. Guardino, J. Chapman, and P. J. Kahrilas, “Ambulatory esophageal pH monitoring using a wireless system,” Amer. J. Gastroenterol., vol. 98, pp. 740–749, 2003.

Jose L. Gonzalez-Guillaumin received the B.Sc. degree in electrical and computer engineering from the Universidad de las Americas Puebla, Puebla, Mexico. He is currently pursuing the Ph.D. degree in electrical engineering at the University of Calgary, Calgary, AB, Canada. His research interests include sensors, instrumentation, imagers, artificial organs, and robotics.

Daniel C. Sadowski received the M.D. degree from the University of Saskatchewan, Saskatoon, SK, Canada, in 1984. He is presently an Associate Clinical Professor of Medicine in the Division of Gastroenterology , University of Alberta, Edmonton, AB, Canada, and the Director of the GI Motility Laboratory at the University of Alberta Hospital, Edmonton, AB, Canada. His research interests are related to clinical methods for assessing gastrointestinal motility disorders. Dr. Sadowski is a certified specialist in Internal Medicine (1990) and Gastroenterology (1991).

Karan V. I. S. Kaler (M’88) received the B.Sc. (Hons.) and Ph.D. degrees in electronic engineering from the University of Wales, Bangor, North Wales, U.K. From 1976 to 1980, he was a Research Fellow at Oklahoma State University, Stillwater, working on cell separation techniques using dielectrophoresis. In 1981, he joined Interalia Associates Ltd., Calgary, AB, Canada, where he was involved in microprocessor-based system design for data acquisition and control purposes. In 1983, he joined the Department of Electrical Engineering, The University of Calgary, Calgary, AB, Canada, where he is presently a Professor. His current research interests are in the application of dielectrophoresis, VLSI, biomedical instrumentation, and electrical stimulation of soft tissue.

Martin P. Mintchev (S’91–M’94–SM’03) received the M.Sc. (Hons.) in electronics from the Technical University of Sofia, Sofia, Bulgaria, in 1987, and the Ph.D. degree in electrical engineering from the University of Alberta, Edmonton, AB, Canada in 1994. In 1994 he completed Post-Doctoral training in Experimental Surgery at the University of Alberta, Edmonton, AB, Canada. Presently, he is a Professor in the Department of Electrical and Computer Engineering, University of Calgary, AB, Canada, and Adjunct Professor of Surgical Research, University of Alberta, Edmonton, AB, Canada. His research interests are related to electronic instrumentation, biomedical engineering and real-time signal processing.