MC

M

F

tapstaactcas

9V

1

icrobubbles during radiofrequency catheter ablation:omposition and formation

ark A. Wood, MD, Katherine M. Shaffer, Amy L. Ellenbogen, Evan D. Ownby, BA

rom Virginia Commonwealth University Medical Center, Richmond, Virginia.

OBJECTIVES The purpose of this study was to measure tissue temperatures associated with micro-bubble formation during radiofrequency (RF) ablation.BACKGROUND Microbubble formation visualized by echocardiography has been used to indicateexcessive tissue heating during RF pulmonary vein isolation. However, little is known about the tissuetemperatures associated with microbubble formation.METHODS Optical fluorometric thermometry probes were used to record tissue temperatures in isolatedporcine atrium overlying either lung or esophageal tissue in a saline bath. RF energy was deliveredthrough an irrigated ablation electrode during echocardiographic monitoring for microbubble forma-tion.RESULTS The maximal recorded tissue temperatures were 81.0 � 5.0°C and 88.3 � 8.1°C at the timeof intermittent (type 1) microbubble formation for lung and esophageal preparations, respectively.During continuous (type 2) microbubble formation, the temperatures were 91.4 � 8.2°C and 99.2 �7.8°C, respectively (both P � .001 vs type 1). Tissue temperatures averaged �100°C at the time of“pops.” The maximal recorded temperature occurred up to 4 mm deep in the tissues and frequentlyoccurred external to the atrial tissue. The total RF lesion volumes for lung and esophageal preparationswere related to the pattern of microbubble formation but not to total power delivered. After generationof type 1 bubbles, up to 60% reductions in RF energy were needed to restore target tissue temperaturesof 65°C. Gas chromatographic analysis of the microbubbles was consistent with steam formation.CONCLUSIONS Microbubble formation during RF ablation represents excessive tissue heating to thepoint of steam formation. Maximal tissue heating may occur in the adjacent lung and esophagus duringcooled ablation.

KEYWORDS Catheter ablation; Atrium; Echocardiography; Lung; Esophagus

(Heart Rhythm 2005;2:397–403) © 2005 Heart Rhythm Society. All rights reserved.

ettbtccT

M

T

fw

Pulmonary vein isolation by radiofrequency (RF) cathe-er ablation is an effective treatment for many patients withtrial fibrillation.1–8 Titration of power delivery near theulmonary veins is important to prevent pulmonary veintenosis.2,7,8 Recently, excessive tissue heating of the pos-erior left atrium has been implicated in fatal cases oftrio-esophageal fistula formation after ablation in the lefttrium.6 Titration of RF energy delivery is difficult withooled ablation electrodes, however, because of the inabilityo monitor tissue temperatures.4,5,9 Under intracardiac echo-ardiography, the visualization of “microbubbles” in the lefttrium during RF ablation has been used to indicate exces-ive tissue temperatures during RF ablation with cooled

Address reprint requests and correspondence: Dr. Mark Wood, Box80053, Virginia Commonwealth University Medical Center, Richmond,irginia 23298-0053.

E-mail address: mwoodmd@pol.net.

n(Received December 5, 2004; accepted December 29, 2004.)

547-5271/$ -see front matter © 2005 Heart Rhythm Society. All rights reserved

lectrodes.1,7 However, little is known about the actualissue temperatures that result in microbubble forma-ion.10,11 The composition of these bubbles and their em-olic risks also are unknown. The purpose of this study waso directly measure tissue temperature, lesion sizes, andharacteristics of RF energy delivery associated with mi-robubble formation in isolated atrial tissue preparations.he composition of the bubbles themselves is examined.

ethods

issue preparation

Pig heart, lung, and esophagus tissues were obtainedrom a local meat processing facility. The right and left atriaere excised and cut into 2 � 2 cm sections with a thick-

ess of 2 mm.12 These atrial sections were placed endocar-

. doi:10.1016/j.hrthm.2004.12.026

dctt4mUsTAPbfvtrewTitdos(

D

cAm4mfe

Riqo

S

iscfapfi

mlepfiIelSb

btbpawtwtt

romeaga1mTtdawre

mFfa

398 Heart Rhythm, Vol 2, No 4, April 2005

ial surface up on a block of either lung tissue (2 � 2 � 2m) or a segment of esophagus (2 cm long � 4–6 mmhick) in a tissue bath filled with normal saline. The bathemperature was maintained at 37°C. The tip of an 8Fr,-mm-tip irrigated ablation catheter (Chilli catheter andodel 8004 RF generator, Boston Scientific, Natick, MA,SA) was positioned perpendicularly to the endocardial

urface of the atrial tissue under 20 gm of contact pressure.he distal electrode was perfused with saline at 36 mL/min.3.5-MHz ultrasound probe (Hewlett Packard Sonos 2500,

alo Alto, CA, USA) was used to visualize the interfaceetween the catheter tip and atrial tissue for microbubbleormation. A pulsatile superfusate flow pattern (mean flowelocity 0.4 m/s) was directed parallel to the atrial tissuehrough plastic tubing (1-cm internal diameter). Four sepa-ate optical fluorometric temperature probes (0.3-mm diam-ter, model STB, Luxtron, Inc., Santa Clara, CA, USA)ere used to record temperatures deep within the tissues.he probes have an accuracy of 0.5°C. Each probe was

ntroduced into the tissue parallel to the ablation catheterhrough separate 22-gauge plastic cannulas inserted to aepth of 8 to 10 mm. The four cannulas were positionedrthogonally about the ablation electrode and cantedlightly so as to pass under the electrode within the tissueFigure 1).

ata acquisition and analysis

All data were recorded on a personal computer usingustom software written in Labview (National Instruments,ustin, TX, USA). Video output from the echocardiographachine was monitored continuously during ablation and

-second video clips captured at the start of each pattern oficrobubble formation (see later). Data from each of the

our temperature probes were sampled at 4 Hz. The ablationlectrode temperature, voltage, and current output from the

igure 1 Schematic of the experimental preparation. A total ofour temperature probes are positioned orthogonally about the

dblation electrode.

F generator were sampled at 30 Hz and the instantaneousmpedance and power calculated at each sample. All ac-uired data and video clips were synchronized in time forffline analysis.

tudy protocols

To examine the composition of the microbubbles, annverted glass funnel devoid of air was submerged in theuperfusate above the tissue block. After 30 minutes ofontinuous RF at 50 W, 0.5 cc of gas accumulated in theunnel. The gas was withdrawn for gas chromatographynalysis by the university’s department of chemistry. Sam-les of gas were generated in both porcine blood and saline-lled baths.

Preliminary experiments demonstrated no difference inaximal tissue temperatures, RF energy, impedance, or

esion sizes for any pattern of bubble formation betweenxperiments conducted in a saline bath compared with ex-eriments performed in a porcine blood-filled bath. Thesendings validated the use of the saline bath for all protocols.n addition, the maximal tissue temperatures recorded atach bubble pattern were examined using an 8-mm-tip ab-ation catheter (EPT Blazer and 1000XP generator, Bostoncientific) and were found to be similar to those generatedy the irrigated-tip catheter.

Three patterns of microbubble formation were definedased on previous descriptions: type 1 microbubbles—in-ermittent, scattered microbubble formation; type 2 micro-ubbles—continuous, dense microbubble formation andop; explosive microbubble formation accompanied by anudible sound.1,7 A fourth pattern, first bubble formation,as defined as the point at which the first bubble was seen

o originate from the electrode-tissue interface. This pointas studied because it was noted to be associated with

issue temperatures of 65 to 70°C, which are near the targetissue temperatures in clinical procedures.9

Four protocols were performed to examine the tissueesponses and energy delivery associated with each patternf microbubble formation. The first protocol examined theaximal temperatures recorded in the tissue at the onset of

ach pattern of microbubble formation (n � 10 experimentstrium over lung and 10 experiments atrium over esopha-us; total 20). The temperature probes were placed about theblation electrode as described. RF energy was delivered at0 W, and the depth of the probes was adjusted such that theaximal tissue temperature was recorded from each probe.his procedure recorded higher maximal tissue tempera-

ures than that of placing the probes at fixed predeterminedepths. The ablation then was stopped and restarted at 25 Wnd increased by 5 W every 20 seconds until 50 W of poweras delivered or a pop occurred. The maximal temperature

ecorded by any of the four probes was noted at the onset ofach pattern of microbubble formation.

The second protocol examined the depth at which theaximal temperature occurred in the tissue preparations

uring each pattern of microbubble formation (n � 10

eoptip

ltm62ttbttdsdl�aoh�Ta

gatbapTpowr6

uT

im

S

wa0csCtutagvcS

R

M

smtcwsbca

M

msaa

T

C

PFTTP

†

399Wood et al Microbubble Formation During Ablation

xperiments atrium over lung and 10 experiments atriumver esophagus; total 20). For this protocol, the four tem-erature probes were set to depths of 2, 4, 6, and 8 mm inhe tissue blocks. RF energy was delivered at 25 W andncreased by 5 W every 20 seconds until 50 W of maximalower was delivered or a pop occurred.

The third protocol examined lesion sizes at energy de-iveries producing different patterns of microbubble forma-ion (n � 10 experiments atrium over lung and 10 experi-ents atrium over esophagus for each bubble pattern; total

0). RF energy was delivered to the tissue blocks starting at5 W, and power was titrated over 20 seconds to producehe first bubble, type 1 or type 2 microbubbles. The powerhen was titrated to maintain the same pattern of micro-ubble formation for an additional 40 seconds. The size ofhe lesions in the atrial and underlying lung or esophagealissues were measured for maximal length, width, andepth. The atrial lesions all were transmural and showedimilar circular to slightly elliptical endocardial and epicar-ial dimensions. The volume of the atrial lesion was calcu-ated as an elliptical solid of uniform thickness: volume �

� r1 � r2 � d, where r1 and r2 are the radii of the longestnd shortest dimensions and d is the thickness. The volumesf the lung and esophageal lesions were calculated as onealf of a prolate spheroid using the formula : volume � 1/2

4/3 � � � d � r2, where r is the radius of the lesion.13

he total lesion volume was calculated as the sum of thetrial and lung or esophageal lesion volumes.

The fourth protocol examined the energy required toenerate a maximal steady-state tissue temperature of 65°Cfter first generating type 1 microbubbles. In clinical prac-ice after observing microbubble formation, RF energy maye titrated downward until bubble formation ceases. Thectual power reduction needed to restore target tissue tem-eratures of approximately 65°C in this setting is unknown.he depths of the temperature probes were optimized as inrotocol 1. RF energy was delivered at 25 W and increasedver 20 seconds to generate type 1 bubbles. The energy thenas decreased over the next 40 seconds until the maximally

ecorded tissue temperature on any of the four probes was5°C.

The velocity of blood flow throughout the atrium is notniform and may influence the results of RF ablation.herefore, all experiments for each protocol were repeated

able 1 Maximal tissue temperatures at microbubble formation

onditionAblation electrodetemperature (°C) Im

reablation 20 � 1 95irst bubble 28 � 3* 75ype 1 29 � 3 72ype 2 31 � 3 71op 36 � 3* 71

*P � .05 vs preceding condition.

P � .05 vs preablation.

n the absence of superfusate flow over the tissue to maxi-ize the effects of flow on the experimental findings.

tatistical analysis

All values are expressed as mean � SD. Comparisonsithin groups were performed using the paired t-test. Power

nalysis indicated that 10 experiments provided a power of.80 to detect a 5°C difference in temperature in pairedomparisons. ANOVA with least squares difference analy-is was used for comparisons among three or more groups.omparisons between independent groups were made using

he independent samples t-test. The Mann-Whitney test wassed for comparison of the tissue depths to maximal lesionemperatures between groups. Linear regression was used tossess variables important to lesion volumes. Ordinal re-ression analysis was used to evaluate the association ofariables with the pattern of bubble formation. P � .05 wasonsidered significant. All analyses were performed usingPSS software (SPSS, Inc., Chicago, IL, USA).

esults

icrobubble composition

The gaseous nature of the microbubbles was demon-trated by the collection of gas from RF delivery. Chro-atographic analysis of the gas showed no difference be-

ween samples generated in saline or blood and no newompounds or change in the proportion of gases comparedith a control sample of air maintained over the saline

uperfusate. The interpretation of these findings was that theubbles represent steam formation and that the water vaporomprising the gas samples equilibrated with the dissolvedir in the saline bath before chromatographic analysis.

icrobubble formation

The maximal tissue temperatures during each pattern oficrobubble formation for both lung and esophageal tissue

amples with superfusate flow are listed in Tables 1 and 2nd Figures 1 and 2. For lung, the maximal tissue temper-ture increased significantly for each pattern of microbubble

/lung

e (�)Power(W)

Maximal tissuetemperature (°C)

0 3726 � 3* 72.1 � 8.8*30 � 2 81.0 � 5.0*38 � 8* 91.4 � 8.2*50 � 6* 105.9 � 7.5*

atrium

pedanc

� 3� 5†� 7†� 6†� 5†

f1cot3tboPm

afeoeppfR.fbaatf

�cotf

cFpdsteiarg(

a1ba

T

C

PFTTP

†‡

Feeapp

Fawa

400 Heart Rhythm, Vol 2, No 4, April 2005

ormation from 72.1 � 8.8°C at the first bubble formation to05.9 � 7.5°C at the time of pops (P � .001 for allomparisons). The ablation electrode temperature averagednly 36 � 3°C at the occurrence of pops and was similar forype 1 and type 2 microbubble formation. Power was 26 �

W at first bubble formation and increased to 50 � 6 W athe time of pop (P � .05 except for type 1 vs first micro-ubble). The impedance decreased significantly from thenset of RF (95 � 3 �) to the first microbubble (75 � 5 �,

� .05) and remained steady for the other patterns oficrobubble formation.For the esophageal tissue, the maximal recorded temper-

ture increased significantly for each pattern of microbubbleormation (P � .001 for all comparisons, Table 2). Thelectrode temperature generally increased for each patternf microbubble formation (P � .05 for all comparisonsxcept type 2 vs type 1 microbubbles). The power increasedrogressively from first bubble formation to the time ofops but was similar between pops and type 2 microbubbleormation (Table 2). The impedance fell from the onset ofF (111 � 8 �) to first bubble formation (71 � 8 �, P �

05) but remained constant for other patterns of microbubbleormation (all P � .05). During type 1 and type 2 micro-ubble formation, the maximal recorded tissue temperaturend ablation electrode temperatures were higher for esoph-geal tissue than for lung tissue (all P � .044). Maximalissue temperature was correlated with the pattern of bubbleormation for both lung and esophageal preparations (both P

able 2 Maximal tissue temperatures at microbubble formation

onditionAblation electrodetemperature (°C) Im

reablation 20 11irst bubble 29 � 3* 7ype 1 34 � 3*‡ 6ype 2 37 � 3‡ 6op 41 � 5* 6

*P � .05 vs preceding condition.P � .05 vs preablation.P � .05 vs atrium/lung.

igure 2 Maximal tissue temperature, radiofrequency (RF) en-rgy impedance, RF power, and ablation electrode temperature atach pattern of microbubble formation for atrium over lung andtrium over esophagus tissue preparations. All experiments wereerformed with superfusate flow over the atrial tissue. *P � .05 vs

receding condition. s

.038) but power, electrode temperature impedance, andhange in impedance were not (all P � .05). The presencer absence of superfusate flow did not significantly alter theissue temperatures at any pattern of microbubble formationor either lung or esophageal preparations (all P � .23).

The depths at which maximal temperatures were re-orded in lung and esophageal tissue are shown in Figure 3.or all patterns of microbubble formation with the lungreparations, the maximal temperature was recorded at aepth of 2 mm from the surface of the atrial tissue withuperfusate flow over the preparation. For the esophagealissue, the maximal temperatures were recorded at depths ofither 2 or 4 mm for all patterns of microbubble formationn the presence of superfusate flow (Figure 3B). In thebsence of superfusate flow, the maximal temperature wasecorded at 4-mm depth in up to 60% of lung and esopha-eal preparations at the time of type 2 microbubbles or popsFigure 3).

The lesion sizes generated in the atrial, lung, and esoph-geal tissues at RF energies producing the first bubble, typemicrobubbles, and type 2 microbubbles are given in Ta-

les 3 and 4. The trends were similar for all groups, with thetrial, lung, esophageal, and total lesion volumes generally

/esophagus

e (�)Power(W)

Maximal tissuetemperature (°C)

0 3725 � 4* 67.3 � 11.8*

‡ 33 � 6* 88.3 � 8.1*‡39 � 7* 99.2 � 7.8*‡

‡ 44 � 5 107.2 � 10.0

igure 3 Depth to maximal recorded tissue temperatures fortrium over lung and atrium over esophagus preparations with andithout superfusate flow over the atrial tissue. The thickness of the

trial tissue was 2 mm in all preparations. *P � .05 vs with

atrium

pedanc

1 � 81 � 8†5 � 7†5 � 7†2 � 5†

uperfusate flow.

gfiub�lalt�

dspata�mia3tttl.

tfrBtta

D

Tbtawtiarap

cng

T

G

FFFNNN

†‡

T

G

FFFNNN

†‡

401Wood et al Microbubble Formation During Ablation

reater with type 2 microbubble formation compared withrst bubble or type 1 bubble formation. Total lesion vol-mes were greater in the absence of superfusate flow foroth lung and esophageal preparations (see Discussion, all P

.05). Transmural esophageal lesions extending to theumen of this structure were commonly noted with type 1nd 2 microbubble formation. By regression analysis, totalesion volumes were related to the pattern of bubble forma-ion (both P � .046) but not total power delivered (both P

.10).After generating type 1 bubbles by titrating RF energy

elivery, the power was decreased to establish a steadytate-maximal tissue temperature of 65°C. The averageowers to generate type 1 microbubbles for lung and esoph-gus preparations were 30 � 2 V and 33 � 6 W, respec-ively. The powers needed to maintain 65°C tissue temper-ture were 19 � 3 W and 19 � 2 W, respectively (both P

.001 vs type 1 bubble formation). The impedance waseasured at the point of power delivery at which the max-

mal tissue temperature first reached 65°C. Upon achievingtissue temperature of 65°C, the impedance fell from 95 �� to 78 � 5 � for lung preparations and from 111 � 7 �

o 68 � 6 � for esophageal preparations (both P � .001). Inhe absence of superfusate flow, the powers needed to main-ain 65°C temperatures were 13 � 2 W and 13 � 4 W forung and esophageal preparations, respectively (both P �001 vs type 1 bubble formation). Upon first achieving

able 3 Lesion sizes atrium/lung preparations

roupAtrial lesionarea (mm2)

Atrial lesionvolume(mm3)

Lung lesioarea (mm2

low first bubble 64 � 26 128 � 51 23 � 14low type 1 72 � 31 144 � 62 41 � 20low type 2 111 � 27 222 � 55†‡ 73 � 12o flow first bubble 90 � 36 194 � 69‡ 42 � 25o flow type 1 101 � 32‡ 254 � 91*‡ 72 � 18‡o flow type 2 161 � 53‡ 357 � 153*‡ 80 � 21

*P � .05 vs first bubble.P � .05 vs type 1 microbubbles.P � .05 flow vs no flow.

able 4 Lesion sizes atrium/esophagus preparations

roupAtrial lesionarea (mm2)

Atrial lesionvolume(mm3)

Esophageallesion area(mm2)

low first bubble 71 � 21 143 � 42 37 � 17low type 1 56 � 24 112 � 47 33 � 16low type 2 99 � 28 198 � 57*† 53 � 16o flow first bubble 84 � 22 168 � 43 57 � 12o flow type 1 94 � 27‡ 188 � 54‡ 58 � 22‡o flow type 2 104 � 29 208 � 58 77 � 13‡

*P � .05 vs first bubble.P � .05 vs type 1 microbubbles.

P � .05 flow vs no flow.

issue temperature of 65°C, the impedance was 84 � 7 �or lung preparations and 80 � 6 � for esophageal prepa-ations (both P � .01 vs preablation) in the absence of flow.y regression analysis, instantaneous RF power, electrode

emperature, impedance, and change of impedance all failedo discriminate between formation of type 1 microbubblesnd maximal tissue temperature of 65°C (all P � .98).

iscussion

he major findings of this study are as follows: (1) micro-ubbles generated during RF ablation indicate excessiveissue temperatures and steam formation; (2) with cooledblation electrodes, maximal tissue heating may occurithin lung or esophageal tissue adjacent to atrial tissue; (3)

he pattern of microbubble formation is predicted by max-mal tissue temperature but not by RF power, impedance, orblation electrode temperature; (4) total RF lesion volume iselated to the pattern of microbubble formation; and (5)fter generation of microbubbles, large reductions in RFower are needed to restore target tissue temperatures.

This study supports the previous assumptions that mi-robubbles represent steam formation.11,14,15,16 The gaseousature of the bubbles was demonstrated, and chromato-raphic analysis did not reveal any new constituents of the

ung lesionepth (mm)

Lung lesionvolume(mm3)

Total lesionvolume (mm3)

Total power(W-s)

.3 � 0.4 31 � 23 159 � 65 977 � 329

.0 � 1.0* 93 � 76 237 � 124 1,464 � 350*

.8 � 1.0*† 206 � 84*† 428 � 113*† 1,609 � 275*

.0 � 1.3 94 � 91‡ 288 � 142‡ 1,357 � 135‡

.3 � 1.5*‡ 252 � 151*‡ 506 � 136*‡ 1,535 � 238*

.7 � 1.2* 293 � 135*† 650 � 169*†‡ 1,747 � 142*†

phagealonth)

Esophageallesionvolume (mm3)

Total lesionvolume(mm3)

Total power(W-s)

� 0.5 74 � 42 217 � 77 1,213 � 284� 0.4 69 � 41 181 � 76 1,597 � 275*� 0.5 121 � 37*† 319 � 83*† 1,776 � 136*� 0.6 133 � 45‡ 301 � 61‡ 1,387 � 118� 1.0 129 � 75‡ 318 � 98‡ 1,803 � 354*� 1.1 223 � 116*†‡ 432 � 144*†‡ 2,088 � 208*†‡

n)

Ld

122233

Esolesidep(mm

1.92.02.32.32.22.8

gtooetbftbt�t0ttsl

wrattct

fosrTcoaoR

ttutfiwmatapdesce

r

tbmtotcpRpf2tii�f

S

siaAlpbetugtm

R

402 Heart Rhythm, Vol 2, No 4, April 2005

as sample compared with control. During the protractedime (30 minutes) necessary to collect the gas, equilibrationf the water vapor and dissolved air in the fluid likelyccurred to produce these findings. This process also wouldxplain the persistence of a gas sample for analysis ratherhan the complete reabsorption of water vapor back into theath. The tissue temperature must be �100°C for steamormation; however, the average maximal tissue tempera-ures were 81°C and 88°C at the onset of type 1 micro-ubble formation for lung and esophageal tissue prepara-ions, respectively. The failure to record temperatures

100°C probably results from the inability to place theemperature probes directly into the very narrow zone (0.5–.8 mm) of active resistive heating about the ablation elec-rode.13 The temperatures were well below 100°C at theime of first bubble formation. This bubble formation pos-ibly represents expansion of gases trapped within the iso-ated tissue sections and not steam formation.

The differences in maximal recorded tissue temperaturesith each pattern of microbubble formation probably rep-

esents rising temperatures in the zone of active heatingnd/or rising heat content of the tissue. The higher recordedemperatures in the esophageal preparations compared withhe lung preparations could be explained by a higher heatonductivity of the dense esophageal tissue compared withhe highly aerated lung.9

Lesion size was correlated with the pattern of bubbleormation but not with total RF power. The poor associationf lesion size with total power is consistent with previoustudies.14,16 The pattern of bubble formation likely directlyeflects tissue heating whereas power delivery does not.14,16

he finding of smaller lesion volumes with superfusate flowompared with no flow is in contrast to the effects of flowver nonirrigated electrodes. Because the irrigated electrodelready is well cooled, it is likely that the superfusate flownly draws heat away from the tissue without augmentingF current delivery.

The maximal heating of tissue deep to the ablation elec-rode is a known consequence of cooled ablation elec-rodes.17 In this study, maximal temperatures were recordedp to 4 mm deep to the endocardial surface of the atrialissue. In patients undergoing catheter ablation for atrialbrillation, the average thickness of the posterior left atrialall is 2.2 � 0.9 mm.18 Thus, for an atrial thickness �4m, the maximal tissue heating may occur outside of the

trial wall. This finding was especially true for esophagealissue and in the absence of superfusate flow over thetrium. The greater density of the esophageal tissue andossibly higher thermal conductivity may contribute to theeeper heating of this tissue. The combination of highersophageal temperatures at microbubble formation and pos-ibly low blood flow velocities in the posterior left atriumompared with the pulmonary vein ostia may make thesophagus particularly vulnerable to thermal injury.

After the appearance of type 1 bubbles, up to 60%

eduction in RF power was needed to restore target tissue

emperatures of 65°C. In clinical practice, RF energy maye titrated upward until microbubbles are seen. Energy thenay be decreased incrementally until microbubble forma-

ion ceases. The findings of our study suggest that reductionf power in small increments may eliminate bubble forma-ion, but the maximal tissue temperature still could be ex-essive at just below 100°C. Complete cessation of RFower may be more prudent to prevent complications. NoF parameter identified the point at which the tissue tem-erature first reached 65°C compared with type 1 bubbleormation. In both cases, the impedance fell dramatically by0 to 30 �, which is generally believed to be excessive ando herald impedance rises in vivo. As with previous studies,mpedance drop itself did not predict excessive tissue heat-ng.14,16 The average ablation electrode temperature was

37°C at the onset of type 1 and type 2 microbubbleormation.

tudy limitations

The studies were performed using isolated porcine tis-ues, and the relevance of this model to the clinical situations not known. No impedance rises occurred in any protocols the result of use of a saline rather than a blood bath.lthough the tissue temperatures were similar between sa-

ine and blood bath preparations, impedance rises may haveroduced different findings at high temperatures in a bloodath system. The study used a stationary tissue sample. Theffects of cardiac motion and different catheter tip orienta-ions may alter the results. The echocardiographic probesed in this study differs from the intracardiac echocardio-raphic system used in the clinical setting.7,15 It is possiblehat differences exist in the ability of these systems to detecticrobubble formation.

eferences

1. Wazni O, Marrouche NF, Martin DO, Gillinov AM, Saliba W, Saad E,Klein A, Bhargava M, Bash D, Schweikert R, Erciyes D, Abdul-KarimA, Brachmann J, Gunther J, Pisano E, Potenza D, Fanelli R, Natale A.Randomized study comparing combined pulmonary vein-left atrialjunction disconnection and cavotricuspid isthmus ablation versus pul-monary vein-left atrial junction disconnection alone in patients pre-senting with typical atrial flutter and atrial fibrillation. Circulation2003;108:2479–2483.

2. Saad EB, Rossillo A, Saad CP, Martin DO, Bhargava M, Erciyes D,Bash D, Williams-Anderews M, Beheiry S, Marrouche NF, Adams J,Pisano E, Fanelli R, Potenza D, Raviele A, Bonso A, ThemistoclakisS, Brachmann J, Saliba WI, Schweikert RA, Natale A. Pulmonary veinstenosis after radiofrequency ablation of atrial fibrillation. Circulation2002;108:3102–3107.

3. Pappone C, Rosanio S, Oreto G, Tocchi M, Gugliotta F, VicedominiG, Salvati A, Dicandia C, Mazzone P, Santinelli V, Gulletta S, Chier-chia S. Circumferential radiofrequency ablation of pulmonary veinostia: a new anatomic approach for curing atrial fibrillation. Circula-tion 2000;102:2619–2628.

4. Oral H, Scharf C, Chugh A, Hall B, Cheung P, Good E, Veerareddy S,

Pelosi F Jr, Morady F. Catheter ablation for paroxysmal atrial fibril-

1

1

1

1

1

1

1

1

1

403Wood et al Microbubble Formation During Ablation

lation: segmental pulmonary vein ostial ablation versus left atrialablation. Circulation 2003;107:1791–1803.

5. Sanders P, Jais P, Hocini M, Hsu L-F, Scavee C, Sacher F, Rotter M,Takahashi Y, Pasquie J-L, Shah DC, Garrigue S, Clementy J, Hais-saguerre M. Electrophysiologic and clinical consequences of linearcatheter ablation to transect the anterior left atrium in patients withatrial fibrillation. Heart Rhythm 2004;1:176–184.

6. Pappone C, Oral H, Santinelli V, Vicedomini G, Lang CC, MangusoF, Torracca L, Benussi S, Alfieri O, Hong R, Lau W, Hirata K,Shikuma N, Hall B, Morady F. Atrio-esophageal fistula as a compli-cation of percutaneous transcatheter ablation of atrial fibrillation. Cir-culation 2004;109:2724–2726.

7. Marrouche NF, Martine DO, Wazni O, Gillinov M, Klein A, BhargavaM, Saad E, Bash D, Yamada H, Jaber W, Schweikert R, Tchou P,Abdul-Karim A, Saliba A, Natale A. Phased-array intracardiac echo-cardiography monitoring during pulmonary vein isolation in patientswith atrial fibrillation: impact on outcome and complications. Circu-lation 2003;107:2710–2716.

8. Purerfellner H, Cihal R, Aichinger J. Pulmonary vein stenosis by ostialirrigated-tip ablation: incidence, time course and prediction. J Cardio-vasc Electrophysiol 2003;14:158–164.

9. Avitall B, Helms R. Determinants of radiofrequency-induced lesionsize. In Huang SKS, Wilber DJ, editors. Radiofrequency CatheterAblation of Cardiac Arrhythmias. Second edition. Futura PublishingCo., Armonk, NY, 2000:47–80.

0. Su WW, Johnson SB, Leite L, Packer DL. The kinetics of tissueheating in pulmonary vein ablation. Pacing Clin Electrophysiol 2003;26:948.

1. Cooper JM, Sapp JL, Tedrow U, Pelligrini CP, Robinson D, Epstein

LM, Stevenson WG. Fire and ice: defining cooled radiofrequency

ablation safety parameters with intracardiac echocardiography. PacingClin Electrophysiol 2003;26:1035.

2. Ho SY, Sanchez-Quintana D, Cabrera JA, Anderson RH. Anatomy ofthe left atrium: Implications for radiofrequency ablation of atrial fi-brillation. J Cardiovasc Electrophysiol 1999;10:1525–1533.

3. Haines DE, Watson DD, Verlow AF. Electrode radius predicts lesionradius during radiofrequency energy heating: validation of a proposedthermodynamic model. Circ Res 1990;67:124–129.

4. Kalman JM, Fitzpatrick AP, Olgin JE, Chin MC, Lee RJ, ScheinmanMM, Lesh MD. Biophysical characteristics of radiofrequency lesionformation in vivo: dynamics of catheter tip-tissue contact evaluated byintracardiac echocardiography. Am Heart J 1997;133:8–18.

5. Martin RE, Ellenbogen KA, Lau YR, Hall JA, Kay GN, Shepard RK,Nixon JV, Wood MA. Phased-array intracardiac echocardiographyduring pulmonary vein isolation and linear ablation for atrial fibrilla-tion. J Cardiovasc Electrophysiol 2002;13:873–879.

6. Saxon LA, Stevenson WG, Fonarow GC, Middlekauf HR, YeatmanLA, Sherman CT, Child JS. Transesophageal echocardiography duringradiofrequency catheter ablation of ventricular tachycardia. Am J Car-diol 1993;72:658–661.

7. Nakagawa H, Yamanashi WS, Pitha JV, Arruda M, Wang X, OhtomoK, Beckman, KJ, McClelland JH, Lazzara R, Jackman WM. Compar-ison of in vivo tissue temperature profile and lesion geometry forradiofrequency ablation with a saline-irrigated electrode versus tem-perature control in a canine thigh muscle preparation. Circulation1995;91:2264–2273.

8. Lemola K, Sneider M, Desjardins B, Case I, Han J, Good E, TamirisaK, Tsemo A, Chugh A, Bogun F, Pelosi F, Kazerooni E, Morady F,Oral H. Computed tomographic analysis of the anatomy of the leftatrium and esophagus. Implications for left atrial catheter ablation.

Circulation 2004;110:3655–3660.