1077

Relation Between Cellular Repolarization Characteristicsand Critical Mass for Human Ventricular Fibrillation

TSU-JUEY WU. M.D., MASAAKI YASHIMA. M.D.. RAHUL DOSHI. M.D..YOUNG-HOON KIM. M.D., CHARLES A. ATHILL. M.D.. JAMES J.C. ONG, M.D.,

LAWRENCE CZER. M.D.. ALFREDO TRENTO. M.D.,*CARLOS BLANCHE. M.D..* ROBERT M. KASS, M.D.,*ALAN GARFINKEL. PH.D..t JAMES N. WEISS, M.D.,t

MICHAEL C. FISHBEIN, M.D.,$ HRAYR S. KARAGUEUZIAN. PH.D.,

and PENG-SHENG CHEN, M.D.

From the Division of Cardiology, Departmenl of Medicine, and *Division of Cardiothoracic Surgery. Department ofSurgery. Cedars-Sinai Medical Center, and the tDivision of Cardiology. Departments of Medicine. Physiology.

and Physiological Science and iJiPathology. UCLA School of Medicine. Los Angeles, California

Critical Mass for Human Ventricular Fibrillation. Introduction: The critical mass forhuman ventricular fibrillation (VF) and its electrical determinants are unclear. The jjoal of thisstudy was to evaluate the relationship between repolarization characteristics and critical massfor VF in diseased human cardiac tissues.

Methods and Results: Fight native hearts from transplant recipients were studied. The rightventricle was immediately excised, then perfused (n = 6) or .superfused (n = 2) with Tyrode'ssolution at 36 C'. The action potential duration (APD) restitution curve was determined by anS)-S2 method. Programmed stimulation and burst pacing were used to induce VF. In 3 of 8tissues, 10 jLiM cromakalim, au ATP-sensitive potassium channel opener, was added to theperfusate and the stimulation protocol repeated. Results show that, at baseline, VF did not occureither spontaneously or during rewarming, and it could not he induced by aggressive electricalstimulation in any tissue. The mean API) at 90% depolarization (APD ,̂,) at a cycle length of 600msec was 227 ± 49 msec, and the mean slope of the API) restitution curve was 0.22 ± 0.08.Among the six tissues perfused, five were not treated with any antiarrhythmic agent. The weightof these five heart samples averaged 111 ± 23 g (range 85 to 138). However, after cromakaliminfusion, sustained VF (> 30 min in duration) was consistently induced. As compared withbaseline in the same tissues, cromakalim shortened the APIX ,̂, from 243 ± 32 msec to 55 ± 18msec (P < 0.001) and increased the maximum slope of the APD restitution curve from 0.24 ±0.11 to 1.43 ± 0.10 (P < 0.01).

Conclusion: At baseline, the critical mass for VF in diseased human hearts in vitro is > 111 g.However, the critical mass for VF can vary, as it can he reduced by shortening APD andincreasing the slope of the APD restitution curve. (J Caniiova.sc Electrophwsiol. Vol. Id pp.1077-1086. Au};u.st 1999)

critical mass, ventricular fibrillation, action potential duration, action potential duration restitutioncune. cromakalim

_ . . . J J • ,. r K,...r-,, ,.• Dr. Wu s curreni address IS Taichung Veterans General Hospital,This study was done durini; the tenure of an NIH Fellowship grant _ . , . _ . '=' 'r* . , -1, r- „ , • , . r , .^ r . . .- - la ichung, Taiwan,to Dr. Athill. a Fellowsnip urant from the Departmenl of Medicine, Z

Korea University. !o Dr. Kim. a Cedars-Sinai FCHO Foundation Address for correspondence: Pcng-Sheni; Chen. M,D,. Division ofAward to Dr. Karagiieu/iaii, a Pauline and Harold Price Kndow- Cardiology, Room 5M2. Cedars-Sinai Medical Center. X700 Bev-nient and an AHA Wyeth-AyersI Established Investigatorship erly Boulevard, Los Angeles, CA 90(M«, I-ax: 310-2894)780; E-Award to Dr. Chen, and was supported in part hy NIH Speciali/.ed mail- chenp@csmc eduCenter for Research (SCORl in Sudden Death Gram P5O-HL.^2319. ^ — '•NIH FIRST Award HL5O2.'>9, and the Ralph M. Parsons Founda- Manuscript received 28 January 1999: Accepted for publication 3tion. Los Angeles. Calitbrnia. May 1999.

1078 Journal of Cardiovascular Electrophysiology Vol. iO. No. 8. Augu.st 1999

Introduction

The generation and maintenance of ventricularfibrillation (VF) depend on many factors. Basedon animal models, these factors may includetissue mass,'"^ cellular electrophysiologic char-acteristics such as action potential duration(APD), conduction velocity,**"̂ and the slope ofthe APD restitution curve.^"^ However, becausethere was no in vitro model of human VF, it wasdifficult to determine the critical mass for humanVF or to test the hypothesis that the critical masscan be modified by perturbation of the electro-physiologic characteristics of cardiac cells. In anattempt to study human VF. we previously dem-onstrated that native hearts harvested from trans-plant recipients readily fibrillated when perfusionof the coronary arteries was restored.'' However,because the si/c of these hearts (494 ± 46 g) wastoo large to be used for in vitro studies, we couldnot record cellular transmembrane potential(TMP) from these hearts to determine their elec-trophysiologic characteristics. We recently de-veloped an isolated swine right ventricle (RV)model in which VF occurred spontaneously dur-ing reperfusion. and TMP recordings could eas-ily be accomplished.^ In this swine model, tissuemass reduction resulted in termination of VFwhen a critical mass (around 19 g) was reached.Encouraged by these results, we sought to deter-mine if a similar perfusion technique could beused to develop a model of VF using the isolateddiseased human RV. In addition, we sought todetermine if altering the repolarization kineticsof human cardiac cells could modify the induc-tion of VF in the isolated human RV.

Materials and Methods

The research protect)! was approved by theInstitutional Review Board of the Cedar.s-SinaiMedical Center. Explanted native hearts fromtransplant recipients were used for this study. Atotal of eight patients were studied. The clinicalcharacteristics of these patients are listed inTable 1.

Tissue Preparation

We selected hearts for this study only whenthe transplant recipient had not previously under-gone open heart surgery. Expiantation of thenative heart was preceded by aortic cross clamp-ing and warm ischemic arrest. Immediately afterexpiantation of the heart, the right coronary ar-tery was cannulated and perfused multiple timeswith cold Tyrode's solution for myocardial pres-ervation. The composition of the Tyrode's solu-tion was reported previously.'° In 2 of 8 patients,the right coronary artery was completely oc-cluded at the orifice, which prevented coronarycannulation. The tissues from these two patientswere superfused for TMP recordings. Theweights of these two tissues were not used forcritical mass analysis.

The RV was excised along with the rightcoronary artery and its orifice. The RV tissue wasimmersed in cold Tyrode's solution and trans-ported to the research laboratory. On arrival, theright coronary artery was immediately cannu-lated and perfused with warm oxygenated Ty-rode's solution at 36°C. The isolated RV wasplaced in the tissue bath. Because the epicardialsurface of the human heart usually is covered by

No,

12*3*4567

Age(Years)

56507557705747^ '

Sexf-MMMMMMM

TABLE 1Clinical Characteristics of Heart

Diagnosis

CADCADCADCADCADCADncMIKAl

LVEF

0.150.180.200.220.170.150.17

Transplant Patients

Clinical .-\rrhyllimi:i.s

VT and VFVT and VFAtrial lihhllationFrequeni PVCsPVCs, syncopeOccasional PVCsOccasional PACsParoxysmal atrial tachycardia

Treatment

A. C. I. ICDA. D. E. F. ICDA, C. F. 1D. F. IC, D. F. ICDC, F, 1D. E. 1D, F

•Tissues with lolal occlusion of the right coronary artery ostium.A = amiodarone; C = captopril; CAD = coronary anery disease; D = digoxin; DCM = dilated cardlomyopathy; E = enalapriU F =furosemide: I = isordil; ICD = implantable cardiovener defibrillator: LVEF = left ventricular ejection fraction; PACs = premature atrialcontractions; PVCs = premature ventricular contractions; VF = ventricular librillation; VT = ventricular tachycardia.

Wi4. et al. Critical Ma.ss for Human Ventricular Kihrillation 1079

a large amount of adipose tissue, the ventricularwail was first placed endocardial side up torTMP recordings. During mapping of activationwavefronts. the tissue was flipped over so thatthe endocardial side was in contact wilh a map-ping electrode array built into the bottom ot thetissue chamber. The tissue bath and the recordingtechniques were the same as those used in aprevious study.' Briefly, a bipolar electrode withan interpolar distance nf 0.5 mm was used torecord bipolar electrograms from the endocar-dium. A pseudo-ECG was registered with widelyspaced bipoles, one at each end of the tissuepreparation. The data, including the TMP record-ings, were acquired by AXON TL-I-4() A/Dacquisition hardware and Axoclamp-2A software(Axon Instruments. Foster City. CA, USA) anddigitized at 1 kHz with 12 bits. Computerizedmapping was performed with the Emap systemmanufactured by Uniservices (Aucktand, NewZealand). The endocardium was constantlypaced with a bipolar electrode at a cycle length of800 nisec for at least 30 minutes before dataacquisition.

Study Protocol

TMP recordings

First, single cell TMP recordings were madewith standard glass tnicroelectrode techniques ineach of the eight tissues.^ When a stable impale-ment was achieved, the refractory period at thehasic pacing (S|) site was determined by theextrastimulus method with twice diastolic thresh-old currents at 5-insec pulse width.'' The longestcoupling interval between S, and the premature.stimulus (S2) that resulted in the noncapture of STwas defined as the effective refractory period.The cycle lengths of S, pacing were 800 and 600msec. The TMP registered during the refractoryperiod testing was used to construct APD resti-tution curves by plotting the APD at 90% repo-larization (APDt),,) on the ordinate and the cor-responding diastolic interval on the abscissa.

Induction of VF

In 6 of the 8 tissues, the right coronary arterieswere patent. To induce VF. a second pair ofbipolar endocardial stimulation electrodes wasplaced 1.5 cm away from the S, site to deliver STin the vulnerable period.'" The S, pacing cycle

length was fixed at 400 msec, and the initialSI-ST coupling interval was 350 msec. The cou-pling interval was progressively reduced in 5- to10-msec increments until loss of capture. Thestrength ol the ST started at 5 mA. If repetitiveactivations were not induced, the strength of theST was increased in 5-mA steps until the induc-tion of reentry or when 50 mA was reached. IfVF was not inducible, additional premature stim-uli were given until a total of three prematurestimuli were used. Burst ventricular pacing wasperformed with progressively shorter cyclelengths and progressively increased currentstrengths. The maximum current strength usedfor burst pacing was 50 mA (four tissues) or IOOniA (two tissues), and the minimum cycle lengthused was 50 msec. At baseline, TMP recordingswere not performed during aggressive electricalstimulation for VF induction.

Cromakatim infusion

In the first 3 of 6 tissues perfused, no VF wasinducible. To test whether or not VF inductionwas dependent on the electrophysiologic charac-teristics of the tissues, 10 /LIM cromakalim wasadded to the perfusate in the remaining threetissues after the baseline study. The same induc-tion protocol was repeated. Cromakalim is anATP-sensitive potassium channel opener'^'"*that significantly shortens APD in ventriculartissues.'^"' Cromakalim was used because of itsknown proarrhythmic effects in the ventri-cles.'^'•''^^ Approximately 15 minutes after be-ginning cromakalim infusion, TMP was regis-tered to determine if there was APD shortening.Attempts were made to induce VF with TMPrecordings. Once VF was induced, endocardialmapping was performed to determine the pat-terns of activation during VF. Reversal of cro-makalim's effect was determined by perfusionwith cromakalim-free Tyrode's solution. As soonas VF terminated. TMP was again recorded, andAPD restitution was reconstructed.

Histologic Examination and Anatomic Correlation

At the conclusion of each experiment, thepreparation was photographed before removalfrom the tissue bath. The specimen was weighedafter removal of the epicardial adipose tissue.The ventricular tissue was fixed in 10% neutralbuffered formalin and processed routinely. Cr<.)sssections were made from the epicardium to the

1080 Journal of Cardiovascular Electrophysiolog> Vol. 10. No. fi. Auf^ust 1999

endocardium and stained with hematoxylin-eosinand trichrome stains. Tissue thickness, myocar-dial fiber orientation, and the presence, if any. oftissue abnormalities were determined.

Data Analysis

The methods for mapping activation duringVF were previously reported in detail.'*''''" Allresults were expressed as mean ± SD. Mann-Whitney tests were used to compare the differ-ences between TMP characteristics in the tissueswith and without amiodarone treatment and toanalyze the effects of cromakalini. Analysis ofvariance was used to compare three groups ofdata. Because the APD restitution curve was flatat baseline, we used linear regression analysis toestimate its slope. After cromakalim infusion,there was a significant increase of slope at shortdiastolic intervals. Linear regression analysiswas applied to the latter part of the restitutioncurve to estimate the maximum slope ofthe APDrestitution curve. P ^ 0.05 was considered sig-nificant.

Results

Table I shows the clinical characteristics ofpatients prior to transplantation. The durationfrom aortic cross clamp to complete excision ofthe hearts ranged from 12 to 21 minutes (mean16 ± 3). The total duration of the experimentsranged from 64 to 210 minutes (mean 1 16 ± 57),during which the pacing threshold remained un-changed. At the end of the experiment, the en-

docardial pacing threshold was 0.4 ± O.I mA.which was similar to the baseline threshold(0.3 ± O.I mA). Visible contractions werepresent in each tissue throughout the experiment.The strength of contraction was diminished bycromakalim infusion, which facilitated the re-cording of TMP during cromakalim washout.

Among the eight hearts studied- two had com-plete occlusion of the right coronary artery ori-fice and cannulation was unsuccessful. Tissuesfrom these two hearts were only used for TMPrecordings during superfusion."' Both of thesetwo tissues and one of tbe remaining six tissueswere from patients treated with amiodarone priorto transplantation. In the remaining five tissues(perfused and not pretreated with antiarrhythmicdrugs), the average weight was 111 ± 23 g(range 85 to 138). This weight was 19.7% ofthemean weight of the entire hearts (563 ± 60 g.range 500 to 635). At baseline, short runs ofrepetitive activations (range 2 to 17 beats) wereinduced by extrastimuli in five tissues: nonedegenerated into VF. Histopathologic studiesshowed interstitial fibrosis and fatty infiltrationsimilar to that reported previously.'̂ There wasno hypereosinophilia. niyocytolysis, autolysis, orother findings of myocardial injury.

Baseline TMP Characteristics andRestitution Curves

Table 2 summarizes the TMP characteristicsmeasured at baseline. APDy,, was measured foraction potentials with an amplitude of at least 60mV (Figs. lA. IB, and 2A), Successful TMP

TABLE 2Transmembranc Poicniial ChanicieriMics in Explanied Heans With or Wiihtnit Aniindaronc Trealmenl

S, PCI, 8(M) msec S, PCI. MM msec

No. of tissues studiedNo. of recording sitesAPDqo (msec)|dV/dtl,n^^ (V/sectAPA (mVlResting membrane

potential (mV|

Amio (

58

251 ±7t ±80 ±

-74 ±

- 1

5934912

Amio

35

282 ±15 ±68 ±

- 6 6 +

( + 1

:57:5:44

NS<0.01<005

NS

Amio (-)

5t l

228 ± 4959 ±3177 ± 7

-71 ± 8

APD Restitution Curve

Amio (-1-1

38

226 ± 5223 ±2268 ± 6

-65 ± 7

P

NS<0.01<0.05

NS

Amui ( - ) Amio ( + )

No. of tissues studiedNo. of APD restitufion furvL'>Slope of APD restitulion curve

47

0.23 r n.lO 0.21 ±0.07 NS

Amio = amiodarone; APA = action poteniial amplitude; APD = action potential duration; PCL = pacing cycte tengih; ( - ) and (+) indicatethe abitence and the presence of amiodarone treatment, respectively.

Wii. et al Critical Mass Ibr Human Ventricular Fibrillation 1081

Figure 1. Transwcmhninc iivtinn potential recording's and action potential duration (APD) restitution curve.s- at baseline. Thedata were from heart 4. Panels A and B show action potential respon.ses to the premature e.xtrastimulus (Sjl methinl for AFDrestitution ciirx'c construction at basic cycle lengths (BCL; S/) of 800 and 6(X) ms. Panel C .v/iovr.v the APD restitution cun-e.The slopes of the restitution curves were < O..i. DI - diastolic interval.

recordings and construction of APD restitutioncurves were obtained in eight and six tissues,respectively. The mean APDy,, at the S, cyclelength of 600 was 227 ± 49 msec (range 135 to294). The mean slope of the APD restitutioncurve was 0.22 ± 0.08 (range 0.07 to 0.35).

In the three patients treated with amiotlaronefor 4, 6, and 10 months before transplantation.there was a reduction of (dV/dt)^,^^ and actionpotential amplitude (Table 2) compatible withamiodarone's sodium cbannel-blocking effect."̂ "̂The slopes of restitution curves averaged 0.21 ±0.07 (range O.I I to 0.29; n = 5) in the tissuespretreated with amiudarone and 0.23 ± 0.10(range 0.07 to 0.35; n = 7) in the tissues withoutamiodarone pretreatment (P = NS). Note that theAPD restitution curve was constructed more thanonce in some tissues (Table 2).

/"Effects of Cromakalim

Cromakalim shortened APDy,, and increasedthe maximum slope of the APD restitution curvein all tissues studied (Table 3). An example is

shown in Figure 2. Cromakalim had no signifi-cant effects un (dV/dt),,,^^, action potential am-plitude, or resting membrane potential. Attemptswere made to induce VF 15 minutes after begin-ning cromakalim infusion. Sustained VF (> 30min in duration) was induced In ali three tissues(Fig. 3). Computerized mapping studies showedthat the VF was characterized by multiple wave-lets similar to those reported for VF in wholehuman hcaits.'*

In all tissues, continuous cromukalim infusionwas associated with sustained VF (> 30 min).Once the cromakalim infusion was stopped. VFeither terminated (n = I) or converted to ven-tricular tachycardia (VT) (n = 2) within 15 min-utes. In the latter two tissues, rapid pacing wasused to terminate VT. TMP recordings wereagain made, and the APD restitution curves werereconstructed (Table 3).

Discussion

A major finding of this study is that VF cannotbe induced in isolated perfused diseased human

1082 .lournal of Cardiava.scular Eleclrophysiology Vol. 10. No. 8. 1999

B CRM, Wash out for 15 min

050 mv

136

..&--..O--O O —-O

Dl < 65 msY=33*1,42'XF.91

100 100 300 100 KM tK

Dl (m;|

• o- : CRM, Wash out for 15 min

°150 mv

1 1 138 99

Finurt' 2. Transmemhranc dciiou potential cluirmtvri.slics at baseline (A) ami liunnii ironuikalim (CRM) infusion (B). Thedata were from heart 5. Panels A ami B show action potential respon.ses to a premature stimulus f5jj before ami aftercnmiakalim infusion. Pam'l C shows APD restitution curves. The maximum .slope of the restitution cutve wit.s increased bycwmakalim. Dl = iliastolic intenxil.

right ventricles thai averaged 111 ± 23 g. or19.7% t)t the mean weight ot the entire hearts. Ina previous study,' we found that explanted nativehearts weighing 494 ± 46 g readily tibrillatedduring repertusion. Based on these findings, thecritical mass lor VF in the native hearts of trans-plant recipients may be at least 111 g but no morethan 494 g. A second major finding is that, ineach tissue. VF became inducible during cro-makalim intusion. The latter finding indicatesthat the critical mass for VF in the diseasedhuman ventricle is not a fixed value, but variesacctjrding to the electrophysiologic state of thecardiac cells.

Critical Mass for VF in Humans and in Animals

The critical mass for VF not only varies ac-cording to the electrophysiologic state of theventricular cells, but it also varies among differ-ent species. The critical mass for VF in canines is25% of the total ventricular weight.' and the

mass of an Isolated canine RV is below thecritical mass for VF."" The critical mass for VFin swine is approximately 16% of the total ven-tricular weight, and the isolated swine RVreadily tibrillates in vitro. In humans, we werenot able to determine the critical mass as a per-centage of the total ventricular weight. However,it is interesting to note that the isolated RVweighing 111 g could not sustain tibrillation.Because III g is many times larger than thecritical mass for VF in either dogs or swine. i(will be interesting to know the mechanism bywhich this difference is observed.

Electrophysiologic Characteristics and theCritical Mass for VF

One possible explanation for the large criticalmass in the diseased human heart relative to thesmall critical mass in healthy canine and swinehearts is that diseased human ventricles have aprok)nged APD.'"^ Increased APD contributes to

Wu, et al. Critical Mass for Human Ventricular Fibrillation 1083

Effects of Cromakalim (10

TABLK 3

on Transmembrane Potential Characteristics of Expianted Hearts

S, PCL 600 msec

No. of tissues studiedNo. of recording sitesAPDy,, (msec)(dV/dt)nH^ (V/sec)APA (mV)Resting nicnibiaiic polcntial (niVi

Ik-fort'

1243 :

71 :7iS -

CRM

t 32t 30f- 7

Duriiif' C

36

55 ±II9±78 ±

-70 ±

APD

RM

1862I I4

ReNlitulion

After

1368975

- 7 0

Curve

( RM

.16::: 21±3R± 9± 6

P*

<0.001NSNSNS

Before CRM During CRM After CRM Pt

No. of tissues sUidiedNo, of APD restiuuion curvesSlope t)f APD rcsiiiiiiion curve 0.24 ±0.11 1.43 ±0.10 <0.01

( - I indicates API) resiitulion curve cannot be ohtained durinji CRM intusion hccuuse pacing ;irKl picmaUire siimulation otlcn induce sustainedventricular arrh>tlimias.*P analyzed by one-wiiy ANOVA. ^̂P anaty/ed by Mann-Whiiney test.CRM = croniakalini; During CRM = CRM int'iisii)n for 15 minutes; After CRM - CRM washout lor 15 niinutes. Other abbrevialinns as inTable 2.

an increa.se of wavelength, which is unfavorablefor the induction of cardiac arrhythmia.^ Fran/et al."^ previously reported a large variation cif

B4 4' 4 i

SI SI S2 S3 $4SOOm

mv1

Figure 3. huimtion of ventricular fibrillation IVF) duringcromakalim Infusion. The data were from heart 5. in whichVF Has not inducible at baseline. After cromakalim infu-sion, sustained VF was easily induced hy triple e.\trastimuli.Panels A and B show simultaneously recorded transmem-brane potentials and bipolar electrograms. These two re-cordings were obtained from different sites on the endocar-diutn. Extremely short action potential dttrations wereobserved. Panel C shows the pseiido-FCd during sustainedVF in the same tissue atquired at a later time.

, measured by monophasic action poten-tial recording techniques in humans. Accordingto Figure 6 of that report, the APD,,H, duringsteady-state pacing at 6tK)-trisec cycle lengthranged from 230 to 300 msec. We also showedthat APDy,, varied significantly from patient topatient. However, the value was consistently> 220 msec al an S| cycle length of 600 msec.In cotnparison, the canitie endocardial APD,,),,mca.sured in our laboratory with the .same tech-niques was around 204 msec at a longer (1.000ms) pacing cycle length.^ Long APD in diseasedhuman ventricles may partially explain the rela-tively large critical tnass observed in this study.

A second possible explanation is the shallowslope of human APD restitution curves. Since thereport by Nolasc() and Dahlen." '̂' numerous stud-ies have demonstrated the itnportance of theslope of the APD testitution curve on the gener-ation of complex catdiac anhythtnia. For exam-ple. Simson et al.'' studied cardiac arrhythmia bycotnputer simulation studies. They reported thatwhen the slope of the APD restitution curve is> 1. the reentrant arrhythmia becomes unstable.A shallow restitution curve, on the other hand,predicts a stable reentrant excitation. Karagueu-zian et al7 studied the APD restitution charac-teristics in canine tissues in vitro and demon-strated that a steep restitution curve is necessaryfor the induction of arrhythmia. Nanasi et al.'̂ ^previously studied the APD restitution curve inisolated ventricular niyocytes from liumans,guinea pigs. dogs, and rabbits at room tempera-

1084 Journal tif Cardiovascular KlectrophysioloKV Vol. 10. No. 8. August 1999

ture. They found that human ventricular cellshave a much more shallow restitution curve thanthose of animal cells. In our study, we also dem-onstrated a very shallow restitution curve in per-fused human RV tissues. This shallow slope ofthe APD restitution curve tnay be responsible forour inability to induce VF.

Cromakalim hifusion resulted in a significantreduction of APD and a significant increase ofthe slope of APD restitution. Electrical stimula-tion resulted in sustained VF. and cromakalimwa.shout resulted in VF termination. These find-Ings also support the idea that, for a given tnassof myocardium, the repolari/ation kinetics con-trol and regulate the inducibility of VF. An in-crease in the slope of the APD restitution curveand a decrease of the APD facilitate the induc-tion of VF. These findings indicate that the crit-ical mass for VF in humans is not a fixed value,but can be modulated according to its reslilutionof repolarization.

ATP-sensitive potassium channel openershave been linked to the development of fibrilla-tion during global ischemia in the perfused ratheart model.'^ However, ischemia also producesmany other electrophysiologic alterations thatare not fully simulated by the opening of theATP-sensitive potassium channels. Because wedid not produce ischemia in our preparation, wewere not able to determine whether or not isch-emia in this preparation can induce similarchanges of APD and restitution curves that facil-itate VF induction.

Study Limitations

A limitation of this study is that these ventric-ular tissues underwent warm ischemia for 16 ± 3minutes before the coronary arteries were per-fused with cold saline ex vivo. Ischemia mayresult in tissue damage and increase the criticalmass for VF; however, various evidence indi-cates that these tissues were stable. (1) Contrac-tions were visible throughout the experiment. (2)The ventricular pacing threshold remained un-changed at the end of each experiment. (3) Dur-ing cromakalim infusion. VF was easily inducedin these tissues; this effect was reversed on wash-out. (4) There was no evidence of autolysis inany tissues studied. (5) The APD measured inthis study was similar to that reported for humanRV endocardial cells in vivo." Furthermore, ourprevious studies of the native hearts of transplantrecipients showed that hearts expianted under

similar conditions fibrillated readily when perfu-sion was resumed.'*

A second limitation of the study is that thesetissues were not contracting against a k>ad. Dueto mechanoelectrical feedback.^^ lack of contrac-tion may result in prolongation of APD,'' pre-venting VF inducibility. However, the tneanAPD.,,, recorded in this study was not signifi-cantly different from the duration of monophasicaction potential recorded in beating humanRVs.^^ The significance of the ab.sence ofmechanoelectrical feedback on the results of ourstudy cannot he determined.

A third limitation is that these hearts were allhypertrophied (mean weight 363 g). The ventri-cles of normal adult hearts usually weigh 350 to400 g. The critical mass for VF may vary accord-ing to underlying structural abnormality and thepresence of hypertrophy. Furthermore, it is wellknown that infants and children can develop VFand die suddenly.""' Whereas the etiology of car-diac arrest in children usually is secondary tononcardiac causes.^' true VF does occur. It isreasonable to assume that some of the pediatrichearts are capable of VF. Therefore, a small massmay not be fully protective against fibrillation,and the 1 1 1-g value may not be applicable to allhuman hearts.

A fourth limitation of the study is that theeffects of cromakalim infusion on APD restitu-tion were determined immediately after cro-makalim washout. It would have been preferableto study APD restitution curve during cro-makalim infusion. However, pacing during cro-makalim infusion often induces VT or VF. mak-ing the determination of APD restitution curveextremely difficult.

Finally, APD and the restitution curves wereall obtained from the RV. Whether or not the leftventricle has the same electrophysiologic char-acteristics is unclear. It is possible that if a dif-ferent part of the ventricle is used, the criticalmass for VF tnay be different.

Clinical Implications

Because the critical mass for VF in humans islarge {> l!l g). at least in this patient popula-tion, it may be possible to use either surgical"^ ornonsurgicai^"' ventricular mass reduction to pre-vent VF. A second clinical implication is that, ifthe critical mass of VF in humans is not a fixedvalue, it may be possible to develop drugs toincrease the critical mass and prevent VF. We

Wu, el al. Critical Mass for Human Ventricular P'ibrillation 1U85

propose that drugs that increase APD and de-crease the slope of tbe APD restitution curvemay be the most effective.

Acknowledgments: The authors thank Duslan Morgan. Avile Mc-Cullen. and Meiling Yuan for their technical assistance and ElaineLebowitz for her secretarial assistance. The authors also wish tothank Peter Hunter, Ph.D.. David Bullivanl. Ph.D.. Sylvain Martel.Ph.D.. and Serge LaFiintaine, Ph.D., for constructing the mappingsystem.

References

1. Garrey WE: The nature of fibrillatory contraction of theheati—Its relation to tissue mass and form. Am JPhysiol 1914;33:397-414,

2. Zipes DP. Fischer J. King RM. et al: Termination ofventricular tihrillation in dog.s by depralarizing a criticalamount of myocardium. Am J Cardioi 1975;36:37-44.

3. Kim Y-H. Gartinkel A. Ikeda T. et al: Spatiotemporalcomplexity of ventricular tibrillatit)n revealed by tissuemass reduction in isolated swine right ventricle. Furtherevidence for the quasiperiodic route to chaos hypothe-sis. J Ciin Invest 1997:100:2486-2500.

4. Moe GK. Rheinboldt WL, Abildskov JA: A computermodel of atrial fibrillation. Am Heart J 1964;64:2OO-220.

5. Rensma PL. Allessie MA, Lammers WJEP, et al:Length of excitation wave and susceptibility to reen-trant atrial arrhythmias in normal conscious dogs. CircRes 198X;62:395-410.

6. Simson MB. Spear JF, Moore EN: Stability of an ex-perimental atrioventricular reentrant tachycardia indogs. Am J Physiol 1981:24():H947-H953.

7. Karagueuzian HS, Khan SS, Hong K, et al: Actionpotential alternans and irregular dynamics in quinidlne-intoxicated ventricular muscle cells. Implications forventricular proarrhythmia. Circulation 1993;87:166I-1672.

8. Garfinkel A. Chen P-S, Walter IX), et al: Quasiperiod-icity and chaos in cardiac fibrillation. J CHn Invest1997:99:305-314.

9. Wu T-J, Ong JJ , Hwang C. et al: Characteristics ofwave fronts during ventricular fibrillation in humanhearts with dilated cardiomyopathy: Role of increasedfibrosis in the generation of reentry. J Am Colt Cardioi1998;32:187-i96,

10. Ikeda T, Uchida T, Hough D, et al: Mechanism ofspontaneous termination of funclional reentry in iso-lated canine right atrium: Evidence for the presence ofan excitable but nonexcited core. Circulation 1996:94:1962-1973.

11. Kobayashi Y, Peters W, Khan SS, et al: Cellular mech-anisms of differential action potential duration restitu-tion in canine ventricular muscle cells during singleversus double premature stimuli. Circulation 1992:86:955-967.

12. Chen P-S, Wolf PD, Dixon EG. et al: Mechanism of

ventricular vulnerability to single premature stimuli inopen-chest dogs. Circ Res 1988;62:1191-1209.

13. Sanguinetti MC. Scott AL, Zingaro GJ, et al: BRL34915 (cromakalim) activates ATP-sensitive K* cur-rent in cardiac muscle. Proc NatI Acad Sci USA 1988;85:8360-8364.

14. Escande D, Thuringer D. Leguern S. et al: The potas-sium channel opener cromakalim (BRL 34915) acti-vates ATP-dependent K' channels in isolated cardiacmyocytes. Biochem Biophys Res Commun 1988:154:620-625.

15. D'Alonzo AJ. Darbenzio RB. Parham CS, et al: Effectsof iniracoronary cromakalim on postischaemic contrac-tile function and action potential duration. CardiovascRes 1992:26:1046-1053.

16. D'Alonzo AJ, Hess TA. Darbenzio RB. et al: Effects ofcromakalim or pinacidil on pacing- and ischemia-induced ventricular fibrillation in the anesthetized pig.Basic Res Cardioi 1994:89:163-176.

17. Wolleben CD. Sanguinetti MC, Siegel PKS: Influenceof ATP-sensitive K^ channel modulators on ischemia-induced fibrillation in isolated rat hearts. J Mol CellCardioi 1989:21:783-788.

18. Kamjoo K. Uchida T, Ikeda T, et al: The importance oflocation and timing of electrical stimuli in terminatingsustained functional reentry in isolated swine ventricu-lar tissues. Evidence in support of a small reentrantcircuit. Circulation 1997:96:2048-2060.

19. Bonometti C, Hwang C. Hough D. et al: Interactionbetween strong electrical stimulation and reentrantwavefronts in canine ventricular fibrillation. Circ Res1995:77:407-416.

20. Lee JJ. Kamjoo K, Hough D, et al: Reentrant wavefronts in Wiggers" stage II ventricular fibrillation: Char-acteristics, and mechanisms of termination and sponta-neous regeneration. Circ Res 1996:78:660-675,

21. Ikeda T, Czer L, Trento A, et al: Induction of mean-dering functional reentrant wavefront in isolated humanatrial tissues. Circulation 1997:96:3013-3020.

22. Sheldon RS, Hill RJ, Cannon NJ. et al: Amiodarone:Biochemical evidence for binding to a receptor for classI drugs associated with the rat cardiac sodium channel.Circ Res 1989:65:477-482,

23. Damiano RJ Jr. Asano T. Smith PK, et al: Effect of theright ventricular isolation procedure on ventricular vul-nerability to fibrillation. J Am Coll Cardioi 1990:15:730-736.

24. Tomaselti GF, Beuckelmann DJ. Calkins HG. et al:Sudden cardiac death in heart failure. The role of ab-normal repolarization. Circulation 1994:90:2534-2539.

25. Franz MR. Swerdlow CD, Liem LB. et al: Cycle lengthdependence of human action potential duration in vivo.Effects of single extrastimuli, sudden sustained rateacceleration and deceleration, and different steady-statefrequencies, J Clin Invest 1988:82:972-979.

26. Nolasco JB. Dahlen RW: A graphic method for thestudy of alternation in cardiac action potentials. J ApplPhysiol 1968;25:191-196.

1086 Journal of Cardiovascular Electrophysiology Vol. 10. No. 8. Aufiiisi 1999

27. Nanasi PP. Varro A. Pankucsi C. cl al: Electrical res-titution in diseased human ventricular my(x:ardium.Clin Physiol 1996;16:339-351.

28. Lab MJ: Contraction-excitation feedback in [nyocar-dium: Physiologic basis and clinical relevance. CirDRes 1982:50:757-766.

29. Chen P-S. Moser KM. Dembitsky WP. et al: Epicardialactivation and repolari/alion patterns in paticnls withright ventricular hypertrophy. Circulation 199I;83:1O4-118.

30. Casado FJ. Mar MF. Valdivielso SA. et al: Fibrila-

cion ventricular en pediatria: tratamiento electrico[Ventricular fibrillation in children: Electrieal treat-ment (author's translation)]. An Esp Pediatr 1979; 12:229-2.34.

31. Graham CA. Scollon D. McGowan J. et al: Resuscita-tion. IV: Advanced paediatric life support. Br J HospMed 1997:58:221-224.

32. Moulopnulos SD. Sidcris DA. Tounianidis ST. et al:induced etectrophysiological isolation of the myocar-dium prevents ventricular fibrillation. Acta Cardiol1991:46:621 630.