Brief Rapid Communications
Observations
on
Electrode-Tissue Interface
Temperature and
Effect
on
Electrical
Impedance
During Radiofrequency Ablation
of
Ventricular
Myocardium
David E. Haines, MD,and AnthonyF. Verow, BA
The purpose of this study was to correlate changes in electrical impedance with the
electrode-tissue interface temperature and to characterize the associated events occurring at
thecathetertipelectrode. In acanine model,lesionswerecreated in vitro (n =49) and invivo (n =31) and radiofrequencypowersettingswerevaried. Electrode-tissueinterfacetemperature, delivered current,andvoltagewererecorded,andimpedancewascalculated. Asudden rise in electricalimpedancewas seeninonlytwoof17 ablations invitro and inoneof 16 ablationsin
vivo withapeak electrodetissue interface temperature of less than100°Ccompared with29 of
32ablations in vitro (p=0.0001)and12of 15ablations in vivo withatemperatureofmorethan 1000 C (p=0.0001). This phenomenon was associated with the observation of boiling and poppingatthetipin invitropreparationsandtissue avulsion and thrombus formationonthe
cathetertipin invivo studies. The lesion sizewasdirectly proportionaltothepeaktemperature for all ablations butnottothepeakpower, current,orvoltage during radiofrequencycatheter
ablation in the heart. Maintaining electrode-tissue interface temperature at less than 1000 C
duringradiofrequency catheter ablation in the heartmay avoid the complicationsassociated with the sudden rise in electrical impedance. (Circukdion 1990;82:1034-1038)
C atheter ablation has beenproposed as a valu-able adjunct in the treatment of sympto-matic sustained tachyarrhythmias.12 Trans-catheter radiofrequency energy ablation causes
myocardialinjury bydirectelectrical
(resistive)
tissueheating aswell as by passive (conductive) heating of
contiguoustissue.3 Although it is relatively safe, the formation of coagulum and thrombus on the
elec-trode surface4 as well as occasional adherent tissue
on catheter removal have been observed with
radio-frequency ablation. Coincident with thesefindings is
asuddenrise inelectricalimpedanceand aresultant
rapid decline in radiofrequency current.4,5 The impedance rise generally occurs at higher levels of delivered radiofrequency power,S6 but the actual causative factor for this phenomenon has not been elucidated.
FromtheDivision ofCardiology, Departmentof Internal
Med-icine, University of Virginia School ofMedicine, Charlottesville,
Va.
Supported by a grant from the American Heart Association, VirginiaAffiliate, Inc., Richmond.
Address forreprints:David E.Haines, MD,Division of Cardi-ology, Box 158, University of Virginia Health Sciences Center, Charlottesville,VA22908.
ReceivedDecember19, 1989; revisionaccepted May 22, 1990.
In the present study, it was hypothesized that a
sudden rise in electrical impedance during
radiofre-quency catheter ablationmight be due toelevation of the electrode-tissue interface temperature above the
boiling point, with resultant boiling of plasma and adherance ofdenatured plasma proteins to the elec-trode. If the electrode-tissue interface temperature is maintained at less than 1000C for the duration of energydelivery, thenboilingwith its associated
imped-ance rise should not occur. Thus, the purpose of this
studywas to assesstheeffectofelectrode-tissue
inter-face temperature during radiofrequencycatheter abla-tiononthe electrical impedanceof the system.
Methods In VitroPreparation
The preparation for an isolated perfused and superfusedcaninerightventricular free wallhas been described in detail elsewhere.3Inbrief, five mongrel dogs (10-35 kg) were anesthetized, and the beating heart was rapidly excised and cooled in iced 0.9%
saline at 40C. The right ventricle with an intact coronary artery pediclewas dissected free from the remainder of the heart and transferred to awarmed
(36.5+0.50 C) tissue chamber. The right coronary arterywascannulated, and the specimenwasperfused
and superfused with Krebs-Henseleit buffer. During
the course ofthe experiment, tissue temperature was continuously monitored. Radiofrequency lesionswere created with a 7F bipolar central lumen electrode catheter attached to afulcrum device thatmaintaineda constant 196 N (20 g) electrode-tissue contact force. In Vivo Preparation
Eight mongrel dogsweighing 10-35kgwere anes-thetizedwithpentobarbitol sodium (30 mg/kg),
intu-bated, and ventilated with a dual-phase control res-pirator pump (Harvard Apparatus, South Natick,
Mass.). Therightfemoralartery wascannulated,and a 7F central lumen bipolar temporary pacemaker
wire (Bard Electrophysiology, Billerica, Mass.) was
positioned within the left ventricle for subsequent radiofrequency lesion production. The catheter posi-tion was documented before each ablation with biplane cinefluoroscopy. After a 2-day recovery
period,theanimalwaseuthanized,andtheheartwas examined grossly and histochemically.
RadiofrequencyLesion Production
All radiofrequency lesions were produced with a Radionics RFG-3AV lesion generator, Burlington, Mass., which produces a continuous 500-kHz sine wave output. For both invitro and in vivo
prepara-tion, the activeelectrode was thetipelectrodeof the
ablation catheter. The indifferent electrode was a 30-cm2 aluminum plateonthefloor of the tissue bath for the invitro
experiments
anda110-cm2cutaneouspatch electrode placed on the left lateral thorax of the in vivo preparation. During lesion production, electrode-tissue interface temperature was
continu-ously monitored with a temperature-sensitive
fiberoptic probe connected to a Luxtron Model 750 Fluoroptic Thermometer, Mountain View, Calif. This probe was advanced through the central lumen ofthe ablation catheter until firm contact with the tissue was made. Radiofrequencypower output was selectedat20% to 100%of the maximum generator output. Actual delivered root mean squared
(RMS)
currentandvoltagewerecontinuouslyrecorded ona
strip-chart recorder. Impedance and RMS power werecalculated at 1-second intervals from themean
measured current and voltage values during that interval. Mean RMS power was the sum of interval power calculations divided by the number of inter-vals. Peak power was defined as the maximal value recorded prior to the observed impedance rise. An
impedance rise was defined as a sudden increase in calculated impedance exceeding a threshold rate of
changeof5% per second. The duration of radiofre-quency energy delivery chosen was 90 seconds because a steady-state tissue temperature rise is
approached at this time.3 Lesion Size
Deternination
Theintact hearts from the in vivoexperimentswere
closely examined, and the lesions were identified by comparing theirgross location with the location of the
TABLE 1. Peak Electrode-Tissue Interface Temperatures and
DeliveredPowerofObservedPatternsofRadiofrequency Ablation
InVitro
Temperature Mean peak Mean peak range temperature power Pattern (0 C) (0 C) (W)
Noboiling 43.6-100.7 81±18 2.2±1.2
Smoothboiling 99.5-113.0 104±5 8.5±5.0
Sudden boiling 99.7-114.3 105±4 17.7±10.8 Popping 102.6-116.8 109±5 19.2±7.8
ablation catheter tip on the cinefluorograms. The lesions wereinspected for evidence of adherent fibrin or thrombus. Lesions from both preparations were bisected and stained with nitro blue tetrazolium
(NBT).7Lesionwidth and depthwere thenrecorded.
Statistics
Rawdata were stored in a computerized database. Categorical data were compared infrequencytables, andsignificance was tested with Fisher's exact statis-tics. Continuous data were presented as mean±SD. Regression lines were determined by least-squares analysis, and goodness-of-fit was determined by anal-ysis of variance and calculation of a multiple corre-lation coefficient.
Results In Vitro Observations
A total of 49 lesions were made with the power output ranging from 10% to 100% and measured mean powers ranging from 0.2 to 52.0 W. During radiofrequency ablation in vitro, four general pat-ternsemerged. Noboilingwas seen atthe electrode
tip in 14 ablations. A pattern of smoothboilingwas observed duringsix ablations as the electrode-tissue interface temperature reachedapproximately1000C.
During 18 ablations, the electrode-tissue interface
temperature achieved or exceeded the temperature ofboiling; then,suddenboilingwasobserved,and the temperature rapidly returned to approximately
1000 C. Finally, 11 ablations followed a pattern
sim-ilar to the sudden boiling group but manifested
audiblepoppingattheelectrode tipwithsubsequent boiling. The peak electrode-tissue interface temper-aturesandmeasured powerdeliveredfor thevarious
patterns arepresented inTable 1.
During 17ofthe 49 ablations, the peak
electrode-tissue interface temperature achieved was less than 1000 C. An impedancerisewas observed inonlytwo ablations (11.8%) in this group (peak temperatures,
99.7°and99.6° C, respectively). Apattern of smooth
orsuddenboilingwasobserved in four ablations(two each)with peaktemperatures ranging from 99.60 to
99.9°C. Of the remaining 13 ablations whose elec-trodetipsdidnotexceed99.5° C,nonedemonstrated
animpedanceriseorvisibleboiling. Thirty-twoof the 49 ablations had a peak electrode-tissue interface temperature ofmorethanorequalto 1000 C. Unlike the former group, a significant impedance rise was
FIGURE 1. Scatterplot of peak electrode-tissue interface tempera-ture recordedduring lesion produc-tionfromall invitro(o) andin vivo
(oi) ablations in experiments with and without a significant rise in electrical impedance. Note strong associationof impedance rise witha
peak temperature exceeding 1000 C.
No Impedance
Rise
measured in 29 of 32 ablations (91%, p=0.0001)
(Figure 1). Only one of 32 ablations showed no boilingatthe electrode tip (peakmeasured
temper-ature, 100.70 C). Four ablationshad smoothboiling, 16 had sudden boiling, and 11 had popping at the
electrode tip. In all cases, the sudden impedance rises observed during ablation in vitro were tempo-rally associated with the observations of boiling or poppingatthe electrode tip.
In Vivo Observations
Atotal of 31 invivo lesionswereproducedwiththe
power output ranging from 35% to 80% and
mea-sured peakpowersranging from 4.8 to 20.0 W.The
peak electrode-tissue interface temperature did not exceed1000 C in16ablations. Ofthose,15 showedno impedance rise. By comparison, 15 ablations had a peaktemperature ofmore than 1000 C (range, 105-1450 C), of which12(80%)showedamarked imped-anceriseranging between91% and 400%persecond
(p=0.0001) (Figure 1).
Examination ofthe electrodecathetersafter lesion
production revealed charring, coagulum, or clot adherenttoallelectrodesinwhichanimpedance rise was observed. In two cases, a small amount of
adherentmyocardiumwasgrossly visible atthe
cath-eter tip. These findings were not observed in cases withoutanimpedancerise. Upongrossinspection of theendocardial surface atnecropsy,fibrinorclotwas adherent to the surface of the lesion in eight of 13 cases where an impedance rise was observed com-pared with two of 18 where there was no rise in impedance (p=0.005).
Univariate analyses were performed to quantify the association between radiofrequency-induced
lesion size and measurements ofenergy delivery to
the myocardium in vivo. No significant relation was seenbetween lesion depth andmean delivered
cur-rent (p=0.38, r=0.16), voltage (p=0.12, r=0.28), power (p=0.63, r=0.09), or peak power before
impedance rise (p=0.28, r=0.20). However, astrong correlationwasfound between both peak and mean electrode-tissue interface temperature, and lesion size (Figure2).
Discussion
Inthepresentstudy, anin vitrosystemof isolated canine myocardiumwas used to directly observe the eventsoccurringatthe electrode-tissue interface dur-ing radiofrequency ablation to elucidate their associ-ation withchanges in electrical impedance. This
pro-tocol demonstrated a close association between sudden boiling, sometimes with audible popping, at
the electrode-tissue interface and a rise in electrical impedance. As one would anticipate, boiling only
occurredwhen theelectrode-tissue interface
temper-atureapproached orexceeded 1000 C. Because anin vitro system with crystalloid superfusate might not
accurately represent in vivo phenomena, a similar protocolwasrepeated inanintact animalpreparation. Intheseexperiments,apowerful associationwasalso
seen between an electrode-tissue interface
tempera-ture exceeding 1000 C and an electrical impedance rise.Itislikely that catheter-tip boiling, similartothat seen in vitro, occurred at the electrode catheter tip whentheboiling pointwasexceeded.
The pathological findings of charring and throm-busfound in thesecasesofferinsight into the mech-anismofthe electricalimpedance rise in vivo during radiofrequency ablation. When the electrode-tissue interface temperature exceeds 1000 C, tissue contig-160 40 . 0 1 120- C-.z 1001 r-o 80' I-s0e... ... -8 oDo 0 o 00 Cl0 003 0 11 0 60t 40 Impedance Rise
*'i-.
p - 0.01
r - 0.75
120 140
FIGURE 2. Scatterplot of lesion
width versus peak electrode-tissue interfacetemperatureforallin vitro
(l)and in vivo (o) lesions.Lesion size from radiofrequency ablation
was directly proportional to peak electrode-tissue interface tempera-ture measured during
radiofre-quency energy delivery irrespective
of the occurrenceofa risein
elec-trical impedance; ablations during
which an impedance rise was
observed are represented by solid symbols.
160
Peak Tissue Temn (1C)
uoustotheelectrode desiccates, andplasma proteins
denaturetoformaninsulating layeronthesurface of the electrode, generally referred to as "coagulum."
Because tissue heating by radiofrequency energy is proportional to the square of current density in tissue,8 a decrease in deliveredcurrent secondaryto
theinsulatingeffects ofcoagulumresults inamarked decrease inheating. In addition, the denatured
pro-teins and dessicated tissueat the site of ablation are probably thrombogenic and create a nidus from which looselyadherent thrombus might embolize.
The ultimate size andshapeof thelesionresulting from radiofrequency ablation may be markedly affected by changes in electrical impedance. Lesion dimension has been demonstrated to be directly proportional to the electrode-tissue interface
tem-peratureduring radiofrequencycatheter ablation.3 It
has been suggested that other parameters of
radio-frequencyenergy delivery, such as currentorpower,
should also be predictive.9 However, the present studydemonstrates thatonce an impedance rise has occurred,the electrical measurements are nolonger useful inpredictinglesionsize.Onlythetemperature
measurementswerepredictiveoflesion size. Because a practical upper limit of radiofrequency heating
existsasthe temperatureapproaches 100° C,a theo-retical maximum lesion size exists for any given electrode size andshape.Forastandard 6Fcatheter, this is estimated to be 7.5 mm in depth. To further increase the size of radiofrequency-induced lesions
andimprovetheefficacyrateofradiofrequency
abla-tion, the final parameters that may be adjusted are the electrode size and surface area. A larger elec-trodetipwill increasethe lesionsize proportional to the electrode radius10 and to the electrode-tissue
interface temperature.
Radiofrequency catheter ablation of the foci of arrhythmiascontinuestobeapromising modalityinthe
treatmentof arrhythmias inavariety of locationswith
varying mechanisms. To optimally control the energy delivery, electrode catheters should incorporate
ther-mometrysothattheamountofradiofrequencyheating
inthemyocardiummaybeaccurately gauged.With this technique, boilingatthecathetertipmaybeprevented.
Thiswill inturnpreventarise in electrical impedance andmaydecrease the chance of thrombus formation.
New catheter designs incorporating larger-tip elec-trodes withtemperature monitors will allow for more predictable and safe but larger and more effective myocardiallesionproduction.
References
1. Scheinman MM,Morady F, HessDS,Gonzalez R: Catheter
induced ablation of the atrioventricular junction to control
refractory supraventricular arrhythmia. JAMA 1982;
248:851-855
2. Critelli G,Perticone F,Coltorti F, MondaV, GallagherJJ:
Antegradeslowbypassconduction after closed-chest ablation of the His bundle in permanent junctional reciprocating tachycardia. Circulation1983;67:687-692
3. Haines DE, Watson DD: Tissue heating during
radiofre-quencycatheterablation: Athermodynamicmodel and obser-vations in isolated perfused and superfused canine right ventricular freewall. PACE1989;12:962-976
4. JackmanWM,KuckK-H,NaccarelliGV,CarmenL,Pitha J:
Radiofrequency current directed across the mitral annulus
with abipolarepicardial-endocardial catheter electrode
con-figurationindogs.Circulation 1988;78:1288-1298
5. Ring ME, Huang SKS, Gorman G, Graham AR:
Determi-nants ofimpedance rise during catheter ablation of bovine
myocardium with radiofrequency energy. PACE 1989;
12:1502-1513
6. Grogan EW Jr, Nellis SH, Subramanian R: Impedance changesduringcatheterablationby radiofrequencyenergy:A
potentialmethod formonitoring efficacyof lesiongeneration (abstract).JAmCollCardiol1987;9:95A
01 o o E Jr. 4J) 10 .94 3C o n 0 0 60 80 100
7. Gottlieb GJ, Kubo SH, Alonso DR: Ultrastructural
character-ization of the border zone surrounding early experimental
myocardial infarcts in dogs.AmJPathol 1981;103:292-303
8. Erez A, Shitzer A: Controlled destruction and temperature
distributions in biological tissues subjected to monoactive electrocoagulation.J Biomech Eng 1980;102:42-49
9. Wittkampf FHIM, Hauer RNW, RoblesdeMedina EO: Con-trol of radiofrequency lesionsizebypowerregulation. Circu-lation1989;80:962-968
10. Haines DE, Watson DD, VerowAF: Electrode radiuspredicts
lesion radius during radiofrequency energyheating.
Valida-tionofaproposed hemodynamic model. Circ Res (in press)
KEY WoRDs * arrhythmias * tachycardia * radiofrequency .
