Gut, 1991,32,240-245

The ion sensitive field effect transistor (ISFET) pHelectrode: a new sensor for long term ambulatory pHmonitoring

Ph Duroux, C Emde, P Bauerfeind, C Francis, A Grisel, L Thybaud, D Armstrong,C Depeursinge, A L Blum

Division ofGastroenterology, CentreHospitalier UniversitaireVaudois, Lausanne,SwitzerlandPh DurouxC EmdeP BauerfeindD ArmstrongA L Blum

Centre Suissed'Electronique et deMicrotechnique,Neuchatel, SwitzerlandC FrancisA Grisel

Institut de PhysiqueAppliquee, EcolePolytechnique FederalLausanne, SwitzerlandL ThybaudC DepeursingeCorrespondence to:Dr Ph Duroux, Division deGastroenterologie, CentreHospitalier UniversitaireVaudois (CHUV/PMU),CH-101 1 Lausanne,Switzerland.

Accepted for publication23 April 1990

AbstractIntraluininal pH monitoring in man should beperformed with disposable multichannelassemblies that allow recordings at multiplesites and prevent transmission of infection.Currently available glass electrodes are un-suitable for this purpose because of their sizeand price. We have thus constructed andtested a small, combined ion sensitive fieldeffect transistor (ISFET) pH electrode in-corporating an integral reference electrode. Invitro studies showed that both ISFET andglass electrodes (440-M4, Ingold, Switzerland)have a linear response over the pH range 1-3-8-0 and that they are comparable with regard toresponse time and 24 hour drift. Twenty onehour intragastric pH recordings were per-formed simultaneously in eight healthy volun-teers using a glass electrode and an ISFETelectrode, placed no more than 2 mm apart in acombined assembly. This was located in thegastric corpus under fluoroscopic control. The21 hour pH curves recorded by each electrodetype showed identical patterns: an early morn-ing rise in pH with three meal-associated pHpeaks lasting for about two to three hours. Themeans of the 21 hour pH medians were 2-09and 2.07 as measured by the glass and theISFET electrodes respectively. Thus, ISFETsare suitable for the construction ofinexpensiveand hence disposable multichannel pHmonitoring assemblies of small diameter. Pro-vided that they can be produced in largenumbers with appropriate technical support,ISFETs have the potential to replace glasselectrodes for long term monitoring of gastro-intestinal luminal acidity.

At present, the standard electrodes for intra-luminal pH monitoring are combined glass elec-trodes which incorporate an integral referenceelectrode,'-3 but although these are highlyaccurate, they have some disadvantages. Studiesof regional pH variations in the upper gastro-intestinal tract require pH assemblies with morethan two channels which, until now, have had touse glass electrodes. These assemblies cannot,however, be passed transnasally as they are toolarge and, if they are passed orally, they cannotbe tolerated for more than a few hours, thusprecluding long term ambulatory recordingswith more than two channels.` Anotherproblem with glass electrodes is that manufac-turing costs are still so high that they are notdisposable. On the other hand, most types ofglass electrodes can only be disinfected: heat and

gas sterilisation and the thermodisinfectionmethod newly developed for fibreopticendoscopes7 are not recommended. Therefore,transmission of infectious diseases such as hepa-titis can never completely be excluded. Becauseof this, the availability of low cost and hencedisposable electrodes is highly desirable.

Attempts to develop a small and inexpensivepH electrode are not new; two examples are themonocrystalline antimony89 and the plastic'0'electrodes. Neither type of electrode, however,is without its problems. Although the behaviourof antimony electrodes in vivo has been reportedto be comparable with that of glass electrodes,'2they are known to be non-linear over a pH rangefrom 1 to 7 and they have a prolonged responsetime compared with glass electrodes.23 Further-more, they require an external reference elec-trode attached to the skin and they may, there-fore, suffer from the fact that the pH electrodeand the reference electrode must work inenvironments that differ in respect of ionicbackground and temperature. Although integralreference electrodes are also susceptible tochanges in the composition of gastric contents, ithas been shown that the use of a distant skin orbuccal reference electrode leads to significantdifferences in recorded pH.2 The main problemwith plastic electrodes is that they cover only aselected pH range depending on membranecomposition and other ions present in thesample. 13 Thus, their output may be predicted invitro when the composition of the measuredsolution is known but not in vivo when, forexample, the intragastric environment changesin the presence of food or refluxed duodenalcontents.We have thus developed a new type of pH

electrode, the ISFET (ion sensitive field effecttransistor) electrode,'"'6 which is theoretically asreliable and accurate as glass electrodes with theadditional advantages that it is smaller and, ifcommercially produced, cheaper. Since com-bined glass pH electrodes represent the goldstandard for 24 hour ambulatory pH monitor-ing,2 we have compared the performance of theISFET electrode with a combined glass electrodeunder in vitro and in vivo conditions.

Methods

ELECTRODES

ISFET electrodeIn principle, an ISFET is a modification of thenormal field effect transistor used in many

240

on 27 May 2018 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.32.3.240 on 1 M

arch 1991. Dow

nloaded from

The ion sensitivefield effect transistor (ISFET)pH electrode: a new sensorfor long term ambulator pH monitoring

Reference electroderl

SaturatedKCI solution

Figure 1: Encapsulation ofan ISFET in a cathetersuitablefor gastrointestinalintubation.CMOS=complementarymetal oxide silicone(integrated circuitconstruction type).

pH -electrode_ s ... .. - -.- ............ - -...... ....... . ...... . \ . . . . ...... ............. : :...... ....... . ...................... .... s...5; ... . .. ....., _ _ .._. ........... .. _ _ .......... _ .... . .

.. _ ..... -... ,_ ,.,.,., :-:.;;r.: .................. ......... ............. .I

CMOS-I1SFET

amplifier circuits. In the ISFET, the metal gate,which is normally used as input, is replaced by anion-sensitive membrane, the measured solution,and a reference electrode. Thus, an ISFETcombines in one device the sensing surface and a

signal amplifier which produces a high current,low impedance output and allows the use ofconnecting cables without excessive shielding.The ISFET used in this study contains a pHsensitive membrane made of A1203.'7 Its ruggedsolid state structure and small size (1.5 x 0-6 x 0 3mm) permitted its encapsulation (Fig 1) in a

polymer cylinder which was placed 4-5 mm fromthe tip of the catheter (1 5 mm outside diameter(OD)) just proximal to the reference electrode.The latter consisted of an AgIAgCl wireimmersed in a saturated KC1 solution behind a

porous glass membrane (CSEM, Neuchatel,Switzerland). A constant-current power supplyis required for the ISFET so that it will produce a

voltage that is linearly proportional to the pH ofthe surrounding fluid over the range ofpH valuesfound in the gastrointestinal tract. The be-haviour of both ISFET and glass electrodes can

be described by the Nernstian equation and thesensitivity of the ISFET electrode is, therefore,comparable with that of a glass pH electrode,although the zero point of the ISFET electrode isdifferent from that of most glass electrodes.

This similarity in behaviour allowed a directcomparison of the electrodes using the same

recording device and evaluation program. Theonly modification required was the addition of a

constant voltage of 1600 mV to the ISFEToutput so that it was in the same range as theoutput of a glass electrode.

Glass electrodeThe combined glass electrode used for validatingthe ISFET is commercially available (440-M4,Dr Ingold AG, Urdorf, Switzerland) and is usedroutinely for upper gastrointestinal pH monitor-ing studies. It has anOD of4mm at the tip oftheelectrode and is mounted on a polyvinyl tubewith an OD of 3 mm. The reference electrode issituated 15 mm proximal to the glass electrode.

in a neutral buffer solution for at least six hoursbefore starting a measurement. The electrodeswere then put in a buffer of pH 4.01 (S1316,Radiometer, Copenhagen, Denmark) at 22°C.The drift exhibited by ISFET electrodes has acharacteristic form with an initial exponentialchange in electrode potential followed by a muchslower but linear drift. In all cases the maximumdeviation from the initial reading was observed atthe end. Therefore, for the assessment of elec-trode drift the voltages were noted at the begin-ning and at the end of a 24 hour test period. Thedrift ofan individual electrode was defined as thedifference in voltage between first and secondmeasurements. The same procedure was used todetermine the.drift of five glass electrodes.

Response timeResponse times were defined as the time takenfor the electrode voltage to reach 90% and 95% ofits final reading in a given neutral or acid buffersolution. The electrodes, which were connectedto a pH meter (PHM 85, Radiometer, Copen-hagen, Denmark) and thence to a chart recorder(REC 80, Radiometer, Copenhagen, Denmark),were transferred between stirred buffer solutionsof pH 7.38 and pH 1-10 at 37°C (S1356 andS 1386, Radiometer). Five electrodes ofeach typewere examined and triple measurements madewith each individual electrode from which meanvalues and ranges were calculated. These data aswell as group medians are presented for eachelectrode type. Differences in response time foreach electrode type were tested statistically usingthe Wilcoxon-Mann-Whitney test.

SensitivityElectrode sensitivity was assessed using fiveelectrodes of each type over the pH range of 1-3to 8. Sodium hydroxide (0.01 N NaOH solution)was added gradually to a stirred solution ofartificial gastric juice at 22°C (NaCl 2 g, Pepsin3.2g, 80ml 1 NHC1 and 1000mlH20,pH 1.3).The electrical potentials of each electrode wereplotted against pH values of the test solution,which were measured with a standard laboratoryglass electrode (GK2401C, Radiometer). Thisprocedure was used to compare the linearity ofresponse of the two electrode types over aphysiological pH range. Linear regression wasused to determine the slope of the data points.

IN VIVO STUDIES

Combined ISFET and glass electrode assemblyFor the in vivo studies of gastric acidity, a

measuring assembly (OD 5 mm) consisting of anISFET electrode and a glass electrode was con-

structed, care being taken that the two sensingsurfaces were never more than 2 mm apart.

IN VITRO STUDIES

DriftEach offive new ISFET electrodes was immersed

Long termpH monitoringEight healthy volunteers (three men and fivewomen, mean age 27 years) participated in thestudy. No subjects received medication, none

were alcohol or drug abusers, and none had anyevidence of gastrointestinal disease. Two volun-teers were smokers; smoking, however, was notallowed during the study. Informed consent wasobtained from all volunteers; the study was

approved by the local ethics committee.Ambulatory 21 hour pH monitoring was con-

ducted in each volunteer, during which theyreceived standard meals. ' No additional food was

1!-

Porous glass /membrane,

,r /

Ag/AgCIwire

74 9 -9

241

on 27 May 2018 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.32.3.240 on 1 M

arch 1991. Dow

nloaded from

Duroux, Emde, Bauerfeind, Francis, Grisel, Thybaud, Armstrong, Depeursinge, Blum

allowed except tap water. The volunteers fol-lowed their usual daily activities.At 4 pm, the combined ISFET/glass electrode

measuring assembly was introduced through thenostril after local anaesthesia and positioned inthe gastric corpus under fluoroscopic control.The electrode cables were connected to separatesolid state data loggers, each with a storagecapacity of 128 kByte (LZ 105, Kaufhold,Berlin, Germany), which were both carried in a

bag along with the box containing the ISFETpower supply and output amplifier.The electrodes were immersed in buffer solu-

tions of pH 7-38 and 1.10 (S1356 and S1386,Radiometer) for five minutes at the beginningand end of each measurement. Electrode poten-tials during these calibration periods, as well as

during the 21 hour intragastric measurement,were sampled at a measuring interval of twoseconds and stored, without transformation topH values, in the data loggers.At the end of each measurement the data were

transferred from the data loggers to a host com-

puter (EuroMak, Dr Weiss GmbH, Schries-heim, Germany). Data were evaluated usingprograms written in the language C and runningunder the OS-9 operating system (Microware,Des Moines, Iowa, USA).

In the initial evaluation step, the first and last30 minutes of the recordings were displayed on

the computer screen, and the operator was askedto mark, manually, the start and end of eachcalibration period. From this the computer thencalculated the electrode zero point and sensi-tivity. These values were used later to transformeach recorded electrode potential into its equiva-lent pH value.

Data presentation and statisticsStandard analysis was performed by calculatingthe median pH for a defined time interval: in thiscase, the full 21 hour recording. In addition,successive median values were calculated from300 consecutive raw data points (correspondingto 10 minute intervals) resulting in 126 datapoints from each ISFET and each glass electroderecording in every subject. These data pointswere used for graphical display and also forstatistical analysis. For each recording, dif-ferences were calculated between the 126 datapairs (glass electrode reading minus ISFETelectrode reading) and plotted against theirmeans'7; bias was expressed as the mean of thesedifferences. As a measure of overall agreement,the median of these mean differences was cal-culated regardless of their sign. Pearson's cor-

relation coefficients were calculated for eachrecording solely in the interest of comparabilitywith previous studies.

Results

IN VITRO STUDIES

DriftThe ISFET electrodes showed drifts of up to15 mV/hour during the first four hours after theirfirst immersion in buffer solution, but after this

200-

175-

125-

100

24 hour drifta

a

KK.

X a

* m

X X:*X -X_ K

0 24Time (hours)

Figure 2: 24 hour drift offive ISFET electrodes (openhourglass symbols) andfive glass electrodes (M).EMF=electromotive force.

stabilisation period there was only minimal longterm drift which never exceeded 15 mV/24 hours(0-26 pH units/24 hours). The 24 hour driftsobtained from all five glass and ISFET electrodesare displayed in Figure 2. Mean drift was7-1 mV/24 hours (0 13 pH units/24 hours) and5.2 mV/24 hours (0.09 pH units/24 hours) forISFET and glass electrodes, respectively.

Response timeISFET and glass electrodes attained 90% of theirfinal reading comparably rapidly after transferfrom acid to neutral buffer but the ISFETsreached 90% of their final reading more rapidlythan the glass electrodes after transfer fromneutral to acid buffer (p<0 01). Glass electrodesattained 95% of their final readings more rapidlythan ISFETs (p<001) in both acid and neutralbuffer solutions (Table I).

SensitivityBoth glass and ISFET electrodes showed a linear

TABLE I Response times ofglass and ISFET electrodes

Neutral to acid Acid to neutral

Buffer 90% 95% 90% 95%

Glass electrode:1 0-87 (0.1) 4-67 (1.7) 0.83 (0.1) 1-70 (0.4)2 1-00(0-2) 3-73(1-5) 0-83(0-1) 1-53(0-1)3 1-07 (0.1) 3.70 (1-0) 0.70 (0.2) 1-30 (0.2)4 0.97 (0.1) 4.10 (1-6) 0.70 (0.2) 1.70 (0.3)5 0.90 (0.2) 3.93 (0.4) 0-80 (0.2) 1*33 (0.1)Median 0 97 3.93 0-80 1.53

ISFET electrode:1 0.53 (0.3) 10-67 (4.2) 0.37 (0 1) 2-37 (1-5)2 0.53(0-1) 10-80(9.0) 0.77(0.6) 2.60(1-2)3 0.73 (0.1) 10-23 (3.1) 0.93 (0.2) 2.13 (0.2)4 0.57 (0.1) 10-67 (3.1) 0.57 (0.3) 2-33 (0.1)5 0-63 (0.4) 8.67 (1.4) 0 83 (0.3) 2.50 (1-0)Median 0.57 10 67 0.77 2-37

Significance * * NS *

The time (seconds) taken to reach 90% and 95% of the final readingis given as the mean (range) of three measurements with eachelectrode and the median value for all five electrodes (*p<0-01:ISFET v glass). Buffers ofpH 7-38 (neutral) and pH 1-10 (acid)were used.

242

on 27 May 2018 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.32.3.240 on 1 M

arch 1991. Dow

nloaded from

The ion sensitivefield effect transistor(ISFE T)pH electrode: a new sensorfor long term ambulatorypH monitoring

5.

pH 4-

3.

2-

1.

I I I ~ ~ ~ ~ ~ ~~~~III I

1800 2100 0000 0300 0600 0900 1200 1500Time (hours)

Figure 3: Mean pH curves ofsimultaneous measurements with ISFET (thick line) and glass(thin line) electrodes in eight healthy volunteers.

response over the pH range from 1 to 8. Themean sensitivity of the glass electrodes was54-5 mV/pH and that ofthe ISFET electrode was57-8 mV/pH.

IN VIVO STUDIESMean pH curves from eight volunteers (Fig 3)show that the recordings from the two electrodetypes are nearly identical. Both show the typicaldiurnal pH variations, with a rise in pH duringthe night and meal-associated pH rises lasting fortwo to three hours at 19:00, 08:00, and 13:00.The means of the individual 21 hour medianswere 2.09 and 2.07 measured with the glass andthe ISFET electrode, respectively.The readings of the two electrodes were com-

pared for each recording by comparing themedian pH values from corresponding, con-secutive 10 minute intervals; the differencesbetween these corresponding pH values are

TABLE II Agreement between electrode readings andcorrelation coefficients

Subject Mean Correlationno difference coefficient

1 0 574 0 9002 0-117 0.9793 0-240 0-8464 -0-0786 0-9605 -0-0282 0-9886 -0-364 0-9787 -0.0512 0.9578 -0-219 0 979Median 0 168* 0-969

Mean difference between electrode readings (glass minus ISFET)and Pearson's correlation coefficient for each subject.* The meandifferences were ranked regardless of their sign to calculate themedian.

plotted against the means of the same two values(Fig 4). These figures indicate that there is not asystematic bias - that is, neither electrode typerecords pH values consistently higher or lowerthan the other. It can also be seen that in most ofthe recordings the differences between the elec-trode readings are generally small; furthermore,these differences are relatively constant and donot increase with increasing gastric pH.The mean difference between electrodes was

greatest in subjects 1 and 6, although this was notreflected by Pearson's correlation coefficient,which indicated a high degree of correlation forall recordings (Table II).

DiscussionAmbulatory long term pH monitoring is now

the main method for evaluating gastric secretoryprofiles and as such it represents an importantresearch tool for investigating the influence ofphysiological'9 and pharmacological factors ongastric acidity. It has also become an essentialclinical tool in the management of gastro-oesophageal reflux disease.20 Upper gastrointes-tinal tract luminal acidity is, however, highly

6 2 4 6 i 0 2 4 6 8

Ii, +'+.4 * .- . .

I . . I

71

. ' t .- -- -1-'t

t ._

1 1- 5 a 1

,6 2 4 6 8 ° 4 6

Average pH using glass and ISFET electrodes

Figure 4: Agreement between ISFET and glass electrodes in each ofthe eight subjects; subjects 1 to 4 are shown on the upper row and subjects 5 to 8 on the lowerrow. The abscissa displays the mean valuefor the two electrode readings; the ordinate the difference between them (glass minus ISFET). The outer horizontal linesrepresent 2 SD either side ofthe mean difference, shown by the central line.

4.1 , -,

2-

0

LLU() -2-nn

~2 -4..I 4-

._0.U

c_ 2-a)1:2._-O3 0

-2.

-4.

243

. t- 1 111 1- ---- -

------ -1

-

~ . '- . r

' C -,.%

9 5 11

a 5 1 5

1

1

-W.. : 'JP 't

Ad-- -,

0 q cr.P-UJ. .'P'

on 27 May 2018 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.32.3.240 on 1 M

arch 1991. Dow

nloaded from

244 Duroux, Emde, Bauerfeind, Francis, Grisel, Thybaud, Armstrong, Depeursinge, Blum

variable with temporal and regional differenceseven within one organ such as the stomach.Thus, pH electrode assemblies with more thantwo channels allowing for simultaneous longterm measurements of pH at different siteswould be highly desirable. In addition, as varia-tions in pH are thought to be dependent on localmotility patterns2' it would also be of greatinterest to use combined pH and pressuremeasurement assemblies for prolonged ambula-tory monitoring. Neither ofthese goals is feasiblewithout the availability of small reliable elec-trodes. Single electrode assemblies with anintegral reference electrode would also be usefulfor localised endoscopic pH measurements as,for example, when investigating Barrett'soesophagus. At present, this can only beachieved by using single glass electrodes with adistant reference electrode. The potentialproblem of disease transmission by non-dis-posable probes, which may be disinfected but aredifficult to sterilise, could also be resolved by theuse of low cost disposable electrodes. Thus, theideal pH electrode should be small, reliable, andcheap while retaining the accuracy and stabilityof glass electrodes, which must, at present, beregarded as the gold standard for pHmonitoring.ISFETs fulfil all these needs. This study

shows that the performance of the ISFET elec-trode is comparable with that of the standardglass electrode in vitro and in vivo. Both types ofelectrode show 90% response times of less than 1second when transferred from neutral to acid andfrom acid to neutral buffer solutions. This allowsan ideal sampling frequency of 1 Hz, which isparticularly important for oesophageal or duo-denal pH monitoring, both of which requirerelatively high sampling rates.22 The responsetime to reach 95% of the final value of the buffersolution is longer with ISFETs than with glasselectrodes but even the response time of thelatter is longer than the 2 second samplinginterval (0 5 Hz) usually used in pH monitoring.The ISFET electrode potential shows a large

initial drift, which is logarithmic in nature,23 butthereafter the drift is negligible. Provided theISFET is placed in a neutral buffer solution atleast four hours before each monitoring, the driftis linear and does not exceed 1 mV/hour in longterm recordings.Median pH values from consecutive 10 minute

intervals were used for the comparison of time-matched values from the individual recordings tominimise the effects of previously documented,rapid gastric pH fluctuations.24 The use of rawpH data values, recorded every 2 seconds, wouldhave overestimated considerably the differencesbetween the electrode types and obscured thefact that they provided comparable measure-ments ofluminal pH over the longer time periodsnormally used in studies of gastric acidity. Thedata show, in.fact, that the differences betweenthe ISFET and glass electrodes in the currentstudy are no greater than those found when twoglass electrodes, in close proximity, are used forsimultaneous gastric pH monitoring.' Further-more, there is no systematic difference betweenthe readings of the two electrode types, and, inthe light of the experience with glass electrodes,

it is probable that many of the differencesbetween the recordings are due to regionalvariations in gastric luminal acidity25"27 ratherthan to differences between the electrodes.

Given that the glass electrode is an ideal pHsensor, the 'worst case' assumption would bethat an ISFET electrode introduces an error of0.168 pH units (Table II) into gastrointestinalpH recordings. This should not have a notableeffect on the required sample size in gastric pHmeasurement studies since even with recordingsystems using a glass electrode, the smallestdetectable difference between test groups oversimilar 24 hour measurement periods was inexcess of 01 pH units.24 Indeed, for shorterperiods, variations in intragastric pH are so largeand rapid that group pH differences of less than1 -5-2-0 pH units cannot be detected.In conclusion, the advent ofISFET electrodes

represents a major advance in intraluminal pHmeasurement. Commercial availability and sup-port will determine whether the ISFET pHelectrode will fulfil its long term potential as areplacement for the glass electrode. In the shortterm, however, it also provides a new techniquefor mapping rapid, localised pH variations whichoccur simultaneously in different parts of thegastrointestinal tract.

This work was presented in part at the annual meeting of theAmerican Gastroenterological Association, May 14-20, 1988,New Orleans. Supported by the Swiss National Foundation grant3.827.0.86.

Further information about the combined ISFET pH electrodeis available from Centre Suisse d'Electronique et de Micro-technique, Neuchatel, Switzerland.

1 Fimmel CJ, Etienne A, Cilluffo T, et al. Long-term ambu-latory gastric pH monitoring: validation of a new methodand effect of H2-antagonists. Gastroenterology 1985; 88:1842-51.

2 McLauchlan G, Rawlings JM, Lucas ML, McCloy RF, CreanGP, McColl KEL. Electrodes for 24-hour pH-monitoring -a comparative study. Gut 1987; 28: 935-9.

3 Emde C, Garner A, Blum AL. Technical aspects of intra-luminal pH-metry in man: current status and recommenda-tions. Gut 1987; 28: 1177-88.

4 Rhodes J, Prestwich CJ. Aciditv at different sites in theproximal duodenum of normal subjects and patients withduodenal ulcer. Gut 1966; 7: 509-20.

5 Rune SJ. In vivo investigations by continuous measurementsof gastric and duodenal pH. ScandJ7 Gastroenterol 1982; 75(suppl): 48-9.

6 Hannibal S, Rune SJ. Duodenal bulb pH in normal subjects.EurJr Clin Invest 1983; 13: 455-60.

7 Working Party of the British Society of Gastroenterology.Cleaning and disinfection of equipment for gastrointestinalflexible endoscopy. Gut 1988; 29: 1134-51.

8 Ask P, Edwall G, Johansson KE, Tibbling L. On the use ofmonocrystalline antimony pH electrodes in gastro-oesophageal functional disorders. Med Biol Eng Comput1982; 20: 383-9.

9 Johansson KE, Tibbling L. Evaluation of the 24-hour pH testat two different levels of the esophagus. In: De Meester TR,Skinner DB, eds. Esophageal disorders: pathophysiology andtherapy. New York: Raven, 1985: 579-82.

10 Rawlings JM, Lucas ML. Plastic pH electrodes for themeasurement of gastrointestinal pH. Gut 1985; 26: 203-7.

11 Oesch U, Brzozka Z, Xu A, et al. Design of neutral hydrogenion carrier for solvent polymeric membrane electrodes ofselected pH-range. Anal Chem 1986; 58: 2285-9.

12 Andersen J, Naesdal J, Strom M. Identical 24-hour gastric pHprofiles when using intragastric antimony or glass electrodesor aspirated gastric juice. Scand Jf Gastroenterol 1988; 23:375-9.

13 Oesch U, Brzozka Z, Xu A, Simon W. Solvent polymericmembrane pH catheter electrode for intraluminal measure-ments in the upper gastrointestinal tract. Med Biol EngComput 1987; 25: 414-9.

14 Bergveld P. Development of an ion-sensitive solid-state devicefor neurophysiology. IEEE Trans Biomed Eng 1970; 19:70-1.

15 Matsuo T, Wise DK. An integrated field-effect electrode forbiopotential recording. IEEE Trans Biomed Eng 1974; 21:485-7.

16 Janata J, Moss SD. Chemically sensitive field-effect tran-sistors. IEEE TransBiomedEng 1976; 21: 241-5.

17 de Rooij NF, Haemmerli A, Grisel A. C-MOS compatiblepH-ISFET with A1203 gate. In: Ko W H, ed. Implantablesensors for closed-loop prosthetic systems. New York: Futura,1985: 139-51.

on 27 May 2018 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.32.3.240 on 1 M

arch 1991. Dow

nloaded from

The ion sensitivefield effect transistor(ISFE T)pH electrode: a new sensorforlongtern ambulatorypH monitoring 245

18 Bland JM, Altman DG. Statistical methods for assessingagreement between two methods of clinical measurement.Lancet 1986; i: 307-10.

19 Duroux Ph, Bauerfeind P, Emde C, Koelz HR, Blum AL.Early dinner reduces nocturnal gastric acidity. Gut 1989; 30:1063-7.

20 Schindlbeck NE, Heinrich C, Konig A, Dendorfer A, Pace F,Muller-Lissner SA. Optimal thresholds, sensitivity andspecificity of long-term pH-metry for the detection ofgastroesophageal reflux disease. Gastroenterology 1987; 93:85-90.

21 Bunun R, Emde C, Bauerfeind P, Stein H, Duroux Ph,Blum AL. Nocturnal gastroesophageal reflux and motility ofthe tubular esophagus: the question of hen and egg[Abstract]. Gastroenterology 1988; 94: A54.

22 Emde C, Bauerfeind P, Bumm R, Blum AL. Minimalsampling rate required in ambulatory long-term pH-metry[Abstract]. DigDisSci 1987; 32: 909.

23 Krawiec J, Odes HS, Schwartz B, Lamprecht SA. Regionaldifference in ambient intraluminal gastric acidity aftercimetidine monitored by intragastric pH-metry.AmJt Gastroenterol 1983; 78: 272-5.

24 Merki HS, Witzel L, Walt RP, et al. Day-to-day variation of24-hour intragastric acidity. Gastroenterology 1988; 94:887-91.

25 Armstrong D, Castiglione F, Biollaz J, et al. Functionalseparation of gastric antrum and corpus observed in gastricdouble pH-metry studies in healthy volunteers [Abstract].Gastroenterology 1989; 96: A15.

26 McLauchlan G, Fullarton GM, Crean GP, McColl KEL.Comparison of gastric body and antral pH: a 24 hourambulatory study in healthy volunteers. Gut 1989; 30:573-8.

27 Schepel SJ, de Rooij NF, Koning G, Oeseburg B, Zijlstra WG.In vivo experiments with a pH-ISFET electrode. Med BiolEng Comput 1984; 22: 6-11.

on 27 May 2018 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.32.3.240 on 1 M

arch 1991. Dow

nloaded from