Arrhythmia Susceptibility and Myocardial Composition in Diabetes Influence of Physical Conditioning Shamsher Bakth, Joseph Arena, William Lee, RemyTorres, Bunyad Haider, BalvantC. Patel, Michael M. Lyons, and Timothy J. Regan Division ofCardiology, Department ofMedicine, UniversityofMedicineandDentistryofNewJersey, NewJerseyMedicalSchool, Newark, NewJersey07103 Abstract We have undertaken a study ofelectrical vulnerability in themyocardiumofacaninemodelafter 1 yrofalloxan-induced Abnormalmyocardialcompositionindiabetesmellitushasbeen diabetes. Since altered cardiac cell action potentials have been described, but the effects on ventricular vulnerability have not described in experimental diabetes with substantial hypergly- beendefined.Wehaveassessed thesusceptibilitytoarrhythmias cemia (5), the action potential characteristics in this model of inacaninemodelafter 1 yrofmilddiabetesinducedbyalloxan. mild diabetes have also been examined. In addition, we have Since physical conditioning can affect metabolic abnormalities assessedtheoccurrenceofspontaneousarrhythmiasduringacute in diabetes, this intervention has also been evaluated. Group 1 myocardial ischemia. servedascontrolsandgroups3and4werediabetic.Animalsin Chronicexercise isknowntoimproveglucosetolerance (6) the lattergroup aswellas nondiabeticcontrols ofgroup 2were andhasbeen showntoincreasethefibrillation thresholdinthe exercised on a treadmill for the last 8 mo of the experiment. basalstateaswellasduringischemiain nondiabetics(7). How- After 1 yr, anesthesia was induced with chloralose for vulnera- ever, it is not known whether the cardiac complications ofdi- bilitystudies. Theventricularfibrillation threshold of24.4±1.9 abetesareaffectedbyenduranceexercise. Intheseexperiments, mAingroup3wassignificantlylessthaninnormals(45.1±2.2). alteredmyocardial composition hasbeenrelatedtoobservations Spontaneous arrhythmias werealsomoreprevalentindiabetics onventricularvulnerability. during acute ischemia (group 3-A). Increased ventricular vul- nerabilityafterepinephrineinfusionwaspresentinthesedentary Methods diabetes despite normal ventricular function responsiveness. In asuperfused preparation ofmyocardium, restingmembrane po- Westudiedhealthymalemongreldogs, 2-4yrofage,weighing22-26 tential andaction potential amplitude werenormal indiabetics, kg.Thedogshadnoclinicalevidenceofdiseaseduringtheinitial8-wk and beta-adrenergic stimulation shortened repolarization more period. Hematocrit, serumalbumin,andelectrolytesaswellasglucose tolerancetestswere normal beforeadmission to the study groups. All thanincontrols. Myocardial collagen concentrations, whichin- theanimals were fed the same diet consistingof8% fat, 22% protein, cludedaninterfibrillar distribution onmorphologic examination, 58%carbohydrate,9%ash,and3%crudefiber.Individualanimalswere were increased in group 3. In the trained diabetics ofgroup 4 placed sequentially in the four groups up to the limit ofthe number thebasal vulnerability thresholdsandresponses toepinephrine plannedforthegivengroup. were normal. While myocardial collagen levels were normal, Forstudiesofventricularvulnerabilityandcomposition,thefollowing cholesterol andtriglyceride increments persisted. Thus, inmild groupswereformed.Group1consistedof11 nonexercisedcontrols,and experimental diabetes, enhanced susceptibility to arrhythmias group2hadfivenormalcontrolsthatwereexercisedfor8mo.Thelatter exists;thissusceptibilitymaybebasedonacombinationofnon- group was limited in size dueto time required for daily exercise ofa homogenouscollagenaccumulationaffectinglocalconductionand totalof14dogs. Diabeticsofgroup3 (n = 9, nonexercised) andgroup increased electrophysiologic sensitivity tocatecholamines. 4(n = 9, exercised) received seriallowdosesofalloxan. Supplemental studiesto furtherdefinethesedentarydiabeticswereperformedinad- ditionalanimalsasindicatedbelowtoassessactionpotentialcharacter- Introduction isticsandarrhythmiasduringischemia. Diabeticmodel.Toproducemilddiabetes,alloxanmonohydratein Abnormalities ofthe myocardium have been described in ani- sterile salinewasadministered intravenously in adoseof20 mg/kgat malswithspontaneous(1)orexperimentaldiabetes(2,3)aswell 3-mointervals. Largerdosesofalloxanwereavoidedtopreventketosis. as in human diabetics (4). Under conditions ofmild diabetes Glucosetolerancewasmeasuredbeforeandat4-mointervalsafterthe without ketoacidosis, modest impairment of left ventricular initialalloxan intheunanesthetizeddogandserialbloodsampleswere function was present (2). Since myocardial composition was taken aspreviouslydescribed (2). Plasmaglucosewasanalyzed bythe found to be abnormal particularly in terms ofcollagen accu- glucose-oxidasemethod(8)tocalculatethedisappearanceratefromthe mulation, the question ofan associated alteration in electrical vascularcompartment. Thisdiabetic model exhibitsreducedglucosetoleranceandplasma propertieshasbeen examined. insulin levels after intravenous glucose (2) as well as abnormal post- prandialglucoseconcentrations(9).Reducedventricularcomplianceand Address reprint requests to Dr. Regan, UMDNJ-New Jersey Medical histochemical changes in the renal mesangium (2) areconsistent with School, 100BergenSt.,Newark, NJ07103. changes in humandiabetes(10, 4). Hemoglobin A1, levels(11)inthis Receivedforpublication 11 February 1985 and in revisedform 6 model were found to risefrom 2.41±0.31% to 3.95±0.33 (P<0.004; September 1985. n = 7)(unpublishedstudies). Astothequestionofadirecttoxiceffectofalloxanonmyocardium, J.Clin. Invest. we have previously observed that myocardial abnormalities were not ©TheAmericanSocietyforClinicalInvestigation, Inc. present when the pancreatic effects ofalloxan and resultant diabetes 0021-9738/86/02/0382/14 $1.00 wereprevented(2).Forthiscurrentstudy,threeanimalsseparatefrom Volume77,February 1986, 382-395 groups 1-4receivedabolusinjectionofglucosetoblockthepancreatic 382 Bakthetal. effects ofalloxan. Ventricular vulnerability was assessed initially and fibrillation,resultinginahighvariance.Inaddition,inrecentpublished after 1 yrwhencollagen contentwasalsodetermined. Thesewerecom- studies,the 5-msdurationhasbeenincommon useindeterminations pared with fourseparate normal controls 9-12 mo after the initial vul- offibrillation thresholdthat do not involve acute ischemia(17). Defi- nerability assessment, which also provided information on long-term brillationwasaccomplishedwithadirectcurrentpulseof50-100W/s, reproducibility. deliveredthrough apairofcopperplates(150cm-2). Thiswasusually Physical conditioning consisted ofrunningon atreadmill (Warren effectivewithin 10-15sanda 1-hintervalwasallowedwhentheanimal E.Collins,Inc.,Braintree, MA)atarateof8mi/honan 180 inclinefor wasretested. 30min/d. Theanimalsexercised 5d/wk for 32 wkafteraninitial 1-2- Afterestablishingthethresholds foreachanimal inthebasalstate, wk period ofgraded exercise (12). Two dogs were excluded from the therepetitiveextrasystolicresponseto l-epinephrine wasassessedsince projectduetoinadequateperformance inthisinitialperiod. Basalheart thishormone isknown toaffect vulnerability(18), andcan bepresent ratewas measured on at leasttwo separate occasions before beginning inenhancedconcentrationsindiabeticsundersomecircumstances(19). thestudyandtoward theend ofthetraining period with thedog lying The hormonewasinfused intravenously ata rate of0.05 to 0.1 tg/kg inaquiet,relaxedstate.Atthesesameperiodsbloodwastakenthrough per min. Metabolic responses were compared by assay ofplasma free intracath tubing for analyses ofplasma lactate (13), before and at the fattyacids(20)andK+concentrations.Attheendoftheabovestudies, termination ofexercise toassessdevelopment oftheconditioned state. clotswereremoved,allvesselsweresutured,andtheskinwasclosed. All dogs were housed in pens large enough to permit free movement, Conduction intervals. Anelectrode catheterwaspassed underfluo- withambulation outside the pensfor 10-15 min/d. roscopy via a carotid artery to the aortic root immediately below the Ventricularvulnerability. Animalsweretestedundersterileconditions valvetorecordaHisbundleelectrogrambeforethevulnerabilitystudies for susceptibility to arrhythmias under chloralose anesthesia, 100 mg/ (21).Bipolarrecordingsweremadewithamultipleelectrodeswitchbox. kg,withsupplementsasneeded. VentilationwasregulatedbyaHarvard TheoutputwasledintotheA-Cinputofanelectrocardiogram pream- pump (Harvard Apparatus Co., Inc., S. Natick, MA) via a cuffed en- plifier, filteredat40-500cps,andrecordedonphotographicpaperata dotrachealtube,and40%oxygenwasusedtopreventhypoxicepisodes. speed of200 mm/s with an Electronics for Medicine recorder (White ArterialpHwasmaintainedwithintherangeof7.36-7.42byadjustment Plains,NY).ThefirstrapidcomponentoftheHisimpulsewastakenas ofrespiratoryrate.Bodytemperaturewasmonitoredwitharectalthermo- itsonset.His-Qtimeshavebeenreproducibleover 1-2hofobservation couple,andwhenbelow370C,wascorrectedbyuseofaninfraredlamp inthedog(21)andduringacuteatrialpacing(22).InadditiontotheP- focusedontheventral surfaceoftheanimal. RandcorrectedQ-Ttimes,theQRSdurationwastakenasthemaximum Alltheanimalsinthefourgroupshadstudiesoftherepetitive ven- observed value in the standard leads. Mean values were obtained by tricularresponsethresholdasanindexofvulnerabilityafter 1 yr.When handtothenearest0.5 msfrom threerepresentativebeatsinthebasal itbecameapparentthatthisvulnerabilityindexdoesnotalwaysparallel state. thefibrillationthreshold(14),determinationsoftheventricularfibrillation Hemodynamics. 2wklater,theleftventricularmechanicalresponse thresholdwerealsoperformed.Thismeasurementwasmadeineightof to catecholamines wasdetermined forcomparison with the prior vul- 11 animalsingroup 1,twooffiveingroup2,sevenofnineingroup3, nerabilitychangesduringarepeatinfusionofl-epinephrineintheanes- andsevenofnineingroup4. thetized state. While vulnerability wastestedin therightventricle, we Inadditiontointergroupcomparisons,controlmeasurementsbefore haveassumedthatthehemodynamic responsecomparedwithcontrols inductionofdiabetesweremadeinsomeofthediabetics.Fouranimals wouldbesimilarinbothventriclesofthediabeticanimals, particularly ofgroup 3 andthree ingroup4wereevaluated beforealloxanadmin- sincethecompositionalchangesvs.controlsdescribedbelowwerecom- istrationand 1 yrlater. parable. Wehavepreviouslyfoundthatthe1-epinephrinedoserangeof Atleast1 helapsedafteranestheticinductionbeforetheformalstudy 0.05-0.1 ag/kgperminelicitsaminimaltomoderatepositiveinotropic began (15). Electrical testing was accomplished with an intracavitary response inthe left ventricle (23). Sincethepeak maximal rate ofrise leadsystem,consistingofatransvenoustripolarcatheter(platinumelec- ofventricularpressure(dP/dt)' intherightventricle, incontrasttothe trodewithan interelectrodedistanceof1 cm andpolewidthof3 mm; left, occursafterthe isovolumic period (24), the use ofdP/dt, isas- N23-7466;Electro-CatheterCorp.,Rahway,NJ)andelectrocardiogram sociatedwithsomewhatgreatererrorintherightascomparedwiththe recording probe (semifloating probe No. 567; Electro-CatheterCorp.). leftventricle.Thus,thelattersitewasconsideredpreferable,particularly Thecathetersystemwaspositionedviajugularveinunderfluoroscopic whentestingatlowintropicdoses. controlattheapexoftherightventricle,withthedistalcathodalelectrode No.8Goodale-Lubincatheters(U.S.CatheterDiv.,C.R.Bard,Inc., localizedinthetrabeculae.Apacemaker(Medtronic,Inc.,Minneapolis, Billerica,MA)wereplacedintheleftventricularchamberandproximal MN)wasemployedtodeliverrectangularpulses, 2 msin duration, to aortafromthecarotidarteriesundersterileconditions. Stathamtrans- maintainaconstantheart rateof200 beats/min, adjustingpacemaker ducerswereused forthemeasurement ofleftventricular(P23Gb)and currenttotwicethemiddiastolicthreshold. Testpulsesweregenerated aortic(P23Db)pressures,whichwererecordedonanamplifier-recorder withanelectricallyisolatedsquare-wavepulsegenerator(S44;GrassIn- system (Electronics forMedicine). Thetransducers were placed atthe strumentCo.,Quincy,MA)andanoperationalpowersupplytoproduce midthoraciclevelandbalancedforequalsensitivity. Themaximalrate constantstimuli. Theelectricaloutputofthisunitwasisolatedandcal- ofriseofleftventricularpressure(dP/dt)wasobtainedusingaresistance- ibrated with an oscilloscope current probe (P6021; Tektronix, Wood- capacitance differentiating circuit. The system had a linear frequency bridge,NJ).Thepulsegeneratorwasequippedwithappropriatecircuitry responsefrom0-30cycles/s.Systolicejectionperiodwasmeasuredfrom toshutofftheoutputofthepacemakerfor3.0safterdeliveryofthetest superimposed aortic and left ventricular pressure pulse tracings. End- stimulus. The test impulse was delivered after every lOth-l5th paced diastolicpressurewasmeasuredoveratleastthreerespiratorycyclesand beat. averaged. These measurements havebeenfoundtocorrelatewellwith A single bipolarstimulus ratherthan atrain ofpulses was used in thepressuremeasurementsobtainedusingamicromanometertip(Millar) viewoftheevidencethatthelatterelicitslocalreleaseofnorepinephrine catheter(25).Asanindexofmyocardialcontractilityweusedthevelocity (16). Electrical scanningwasconducted at 3 ms intervalsbeginningat ofcontractile element (Vce) at peak dP/dt, calculated as the ratio of the end ofthe effective refractory period and terminating 10 ms after maximaldP/dtrelatedtothesimultaneousleftventricularpressure(26). theT-wave.Usingelectricpulsesof2msdurationthestimulusintensity was increased in 2 mA steps and scanning continued until repetitive extrasystoles were elicited. The lowest stimulus intensity that evoked 1. Abbreviations usedin thispaper: APA, action potential amplitude; repetitiveextrasystolicbeatsintwoofthreetrialsdefinedthisthreshold. APD5o, actionpotentialdurationat50%repolarization;APD80,action Todeterminetheventricularfibrillationthresholdweusedateststimulus potentialduration at80%repolarization; dPdt, maximal rateofriseof durationof5 ms,sincetheshorterpulsedurationinpreliminarystudies ventricularpressure;MDP, maximumdiastolicpotential;Vce,velocity ofnormalsoccasionallyrequiredan unusuallyhighcurrenttoproduce ofcontractileelement. Diabetes, CardiacComposition, andVulnerability 383 Myocardial composition. 2 wk were allowed for recovery and the endocardialtissuewaspinnedwithsurfaceupina30-mlchamberand animals were reanesthetized. 1 h later the thorax was incised and the superfusedat 10ml/minofTyrode'sequilibratedwith95%02/5%CO2. heartrapidlyarrestedwithiced Ringer'ssolution. Transmural samples Temperaturewasmaintainedat37.0±0.50CandthepHat7.3to7.4. from both ventricles were placed in liquid nitrogen forglycogen assay Theisolatedmyocardiumwasstimulatedthroughbipolarsilver-wire (27).Bothventriclesandseptumwereseparatelyweighed.Sampleswere electrodesatarateof60min-'beginning30minaftersuperfusionwas takenfromtheinnerandouterhalvesofleftventricle,theseptum,and begun.Thestimulatingelectrodewaslocated5-10mmfromtherecording rightventricleforhydroxyprolineanalysestoestimatecollagenconcen- electrode. Rectangularpulses, 1-5 msinduration andtwicethreshold tration(28).Transmuralsamplesofleftventriclewereextractedforassay voltage,weregeneratedbyastimulator(GrassInstrumentCo.)isolated ofcholesterol (29), phospholipid (30), andtriglyceride (31). Forcation fromground.Transmembraneactionpotentialswererecordedthrough analysis the samples were homogenized and extracted for 48-72 h in glasscapillary microelectrodesfilledwith3MKCIatatipresistanceof distilled water, sufficient time forcomplete extraction. Potassium and 5-15megohms.Microelectrodeswerecoupledbyasilver-silverchloride sodium were determined on an Autoanalyzer system (Technicon In- wiretoahighinputimpedanceamplifierwithcapacitanceneutralization. strumentsCorp.,Tarrytown,NY)withflameattachment.Wedetermined Thedepthofthefluidoverthepreparationwasmaintainedataminimal water content ofthe myocardium by drying the tissue samples in an leveltoreducecapacitanceintherecordingsystem. Theoutputofthe ovenat 1IO'Ctoconstantweight, amplifierwasdisplayed on an oscilloscope and photographed with an Forahistochemical assessment, transmural sampleswere taken as oscilloscopecamera(C4;Grass Instrument Co.). Action potential am- unknownsfromthemid-leftventricleofthreeanimalspergroup.These plitude(APA),maximumdiastolicpotential(MDP),andactionpotential examinations included trichrome stain for connective tissue (32, 33). durationat50%(APDjo)and80%repolarization(APD80)wereanalyzed Forelectron microscopy, transmuralsampleswerefixedina2.5%cold from this film. After an initial 30-min equilibration, action potential glutaraldehyde bufferedwithphosphate, washed, postfixed in osmium, characteristicsweredocumentedbyrecordingPurkinjetypeactionpo- exposedtoleadanduranylacetate, andembeddedinepon. Eight l- m tentialsfromtheendocardialsurfaceandventricularmusclepotentials sections were prepared for each dog. Approximately ten micrographs frombeneaththesubendocardial layers;7-15 cellsweresampled from wereobtainedfromeachsectiontoacquirealongitudinalorientationat eachanimal. anappropriate magnification. Aftercontrol recordings, incrementaldosesoffreshlypreparediso- Acutemyocardial ischemia. Aftercompletion oftheabovestudies, proterenolweresuperfused, 30minforeachdose,atwhichtimesteady weproposedtodeterminewhetherspontaneousarrhythmiasduringacute state measurements were taken. Doses ranged from 10-8 to l0-s M. myocardialischemiahadadifferentprevalenceindiabeticanimalsthan Subsequently,drug-freesolutionwassuperfusedtocheckforreversibility controls. Eight male mongrel dogs received alloxan, as above. These ofisoproterenoleffects. sedentary diabetics (group 3-A) were compared with normal controls Statistics. Thedataareexpressedasmean±SE.Whenonlyonesta- (group 1-A), matchedforageandsex. tisticalcomparison wasemployedbetween acontrolandintervention, After 11-12 mo the dogs were anesthetized with chloralose. Oxy- Student'sttestforpaireddatawasused.Toassesssignificancebetween genation,bodytemperature,andbloodpHweremaintainedasdescribed groups,analysisofvariancewasperformedusingDuncan'smultiplerange above.Withchestintact,adoublelumen'No. 5balloon-tippedcatheter test for comparisons when F values were significant, at an a level of was positioned via the carotid artery. Coronary angiography was per- <0.05 (38). Fisher'sexacttestwasusedtoevaluatearrhythmiasinthe forme4beforeischemiatodeterminethedistributionandcourseofthe acute myocardial ischemia study. Action potential duration responses leftanteriordescendingcoronaryartery. Two normal dogsandonedi- togradeddosesofisoproterenolwerecomparedbyattestfordifferences abeticdogwereexcludedbecausetheanteriordescendingarterydidnot inregressionlineelevation(38). coursetotheapexoramajorobliquevesseloriginatedclosetotheorigin oftheanteriordescendens,thusdiminishingtheareaatrisk.Afterballoon Results inflationintheproximal 1.5cmoftheleftanteriordescendingcoronary artery(34),transmural ischemiawasevidencedbyareductionofmean Bodyweightinanimalsofthefourgroupswasmaintainedover peripheralcoronaryarterypressureto -25mmHgandtheappearance the 12moofobservationwhileplasmaalbuminandelectrolytes ofaninjury potential onstandardlead 1. Nocardioactive agentswere remainedwithinthenormalrange(TableI).Diabeticsofgroups given. Theanimalswereobservedupto 15min,ariskperiodforarrhyth- 3and4exhibitedareductionofglucosetolerancethatwassimilar mias.Thislimitwasimposedsincethisdiabeticmodelexhibitsagreater inextentat4moaftertheinitialalloxanwithaGkof1.91±0.27 degreeofleftventriculardysfunctionduringischemiaovertimethando and 1.82±0.14, respectively. Fasting normoglycemia was not normals(34),whichmayrepresentanadditionalvariableaffectingout- affected by the exercise regimen in group 4 diabetics, but at 1 come.Arrhythmiaswererecordedata25mm/spaperspeedonarecorder yrtheglucoseclearanceconstantwasincreasedto2.40±0.15 (P (ElectronicsforMedicine)withautomaticdevel6per. <0.05), without achange ingroup 3. Ventriculartachycardia wasdefinedas'aseriesofsixormore con- Evaluation ofthemodels. Toexaminethepotentialcardio- secutive ventricularectopiccomplexes(35). Weexaminedthetotal in- toxiceffectsofalloxan, wepreventedthedevelopmentofglucose cidence ofventriculartachycardia andfibrillation inthetwogroupsas intoleranceinthreeanimalsafterpretreatmentwith50gofglu- wellasthesumofventricularectopiccomplexes(35, 36).Inthosethat survived 15min,theheartwascold-arrestedwithicedRinger'ssolution, cose before each alloxan dosage. The fibrillation thresholdwas andEvan'sBluedyewassimultaneouslyadministeredviathecoronary unchangedafter 1 yrintheseanimalsaswellasinfournormals arterycatheterataortic-diastolicpressurelevelstodeterminetheareaat followed for 9-12 months without such treatment (Table II). risk. In animals that fibrillated, the injection pressure wasthat ofthe Collagenconcentrations intransmural samplesofleftventricle, residualaorticpressure. as a marker of diabetic changes, were 2.14±0.19 ag/mg dry Electrophysiologicalstudiesinvitro.Todeterminewhetherthismodel weight in the glucose-alloxan group vs. 2.09±0.11 in normals. ofdiabeteswasassociatedwithalteredcellactionpotentials,werecorded Thus, acardiotoxic effectofthiscompound appears unlikely. intracellularpotentials(37), andtheresponse ofthesepotentialstoiso- Adaptation to chronic exercise was indicated in groups 2 proterenolinfournormals(group 1-C)andthreediabetics(group3-C) and 4 by a reduction ofbasal heart rate in the conscious state preparedasabove. After 1 yrtheheartwascold-arrested, andtrans- and the failure ofblood lactate to rise during acute exercise muralsampleswererapidlytakenfromthefreewalloftheleftventricle 2.5cmabovetheapexandplacedincooloxygenatedphysiologicTyrode's (Table I). In addition, left ventricular weight, as a measure of solution with the followingmillimolar composition: NaCI, 135; KCl, hypertrophy, increased similarly in these two groups. 4.0; dextrose, 5; CaCl2, 2.0; MgCl2, 1.0; NaHCO3, 15; Na2HPO4, 1.0. Ventricularvulnerabilityindiabetics. Incontrolanimalsfor After removal ofthe epicardial two-thirds ofthe ventricular slice, the thediabeticseries(group 1),therepetitiveextrasystolicthreshold 384 Bakthetal. 00 -(i4-lCi_ car co %- +o1o+1 o+1 +o1 +I+1++30-C0I o+1~ wL .e t .* .qe .o 00 r 0% .0- 0 +-°H +1 + +1+ 0. u %O en t I! .t4t- 0 t_ 4) -1en I ^A t C) 1 .c9o : 0+0-+100+1 +1 4)-4) 0 2 04 _)C +1++1O+11%O++1 u Lu +1 +1 +1 +1 c0%r'000o0 en .I _. _ e_n = = 5:aL +l +l +l +l ten %o en +l +1 +1 +1 en _ en E _q _R I'_ _~ - %o ,- 0 t I) 0 0 9 _9-~H+_1+1+o1 wL COD CAM +1 +1 +1 wL .n i,een I+t1 +t1 I+R +1V+Rt 0* o o~ o O r- 0l00C- I o V e+1eo4i+1-v+_i10_T+l IR X: +l +l +Tl +l V u 0r-0 Cr-o. o0o o0o1 .;4OEZ ,e-_ e-4o0O-- . . . 0 +l +l +l+ ~< U Cuc) u 0.)) 0en0 Cr-m4 C0-~.-0 o+ol t+_1a- +o1o om 8 a+s1+-o10+0e,14 0+1 C'r 8 wL IvRi -l C--i Cus _ V00) 6O+o+lo+1 0+%1 %cOoo0rCO1- O1n o 40u 00 00 o% - o *0;_03 u . C- -D o.. 10 r- ) .0 0+%1 0+01 %-.H0 e+n1 .GC~- C0+-J40C-+--lI C0+o-0li Oe+1Nl U Itt 0-0 er- r0q% 00 3 +1 -~ k .9 e0+0nl +0C-01-+l0'0+0* 0~+o10o- CZ N 00 1- - eI4In OI _L I1I1 11 11 1I.-, 00 D00 z_.St_ I_ 0 Diabetes, CardiacComposition, and Vulnerability 385 TableII. Serial VentricularFibrillation ThresholdinNormals* TableIII. VentricularFibrillation Threshold* Animalno. Initial 12mo Group 1 Group3 Group4 mA mA mA mA mA 1 48* 46 38 27 41 2 53* 55 42 31 38 3 42* 41 44 18 43 4 39 43 45 22 36 5 45 39 43 20 37 6 38 35 48 31 46 7 49 48 49 22 48 Mean±SE 44.85±2.09 43.86±2.47 52 45.1±1.6 24.4±1.9 41.3±1.7 * Indicatesanimalsthatreceivedglucose-alloxan aftertheinitialdeter- minationoffibrillation threshold. Theotherfouranimalsdidnotre- * Threshold forindividualanimalsasdeterminedaftertherepetitive ceiveglucoseoralloxan. ventricularresponse. Unpairedttest:group 1 vs. 3,P<0.005;group 1 vs.4,N.S.;group 3 vs.4,P< 0.01. was 45.8±2.0 mA in the basal state after 1 yr (Fig. 1). The threshold wassignificantly lowerinthe nontrained diabetics of group 3 at a level of27.5±3.8 mA. By contrast, in the trained betweencontrolsanddiabetics(TableV). Basalplasmafreefatty diabetics ofgroup 4, this threshold at 12 mo was significantly acid(FFA)levelswerealsosimilarbetweengroup 1 andthetwo higherthaninthesedentarydiabetics,didnotdifferfromgroup diabetic groups, with similar elevations during the infusion of 1 controls,butwaslessthanthelevelintheconditioned normals epinephrine. Group 2 controls differed from group 1 with a (Fig. 1). higher basal FFA level. Neither plasma K+ decreases nor ele- Evaluationoftheventricularfibrillationthresholdinthebasal vationofplasmaFFAappearedtocontributetothevulnerability staterevealedasignificantreductioninthenontraineddiabetics differences in untrained andtraineddiabetics. ofgroup3comparedwithcontrolsofgroup 1 (TableIII). Inthe Duringthe hemodynamic study 2 wk later, left ventricular traineddiabeticsofgroup4thethresholdwassignificantlyhigher performance during administration ofepinephrine was com- than that ofthe sedentary diabetics and approximated that of paredintermsofvelocityofcontractileelement(Vce)andfilling thenormalcontrols. Onlytwooftheexercisednormalsofgroup pressure (Table V). The responses in diabetics ofgroups 3 and 2 weretested and hadfibrillation thresholds of51 and 53 mA. 4did notdiffer from those ofthe normals. Inthe subset ofgroup 3 animalswith measurements before Conduction times. In thebasal state, P-R andQRS on lead induction ofdiabetes aswell as 1 yrlater, the extrasystolic and fibrillationthresholdsweresignificantly reducedattheterminal studies (Table IV). The group 4 subset showed no change in TableIV. Serial Vulnerability Thresholds eitherparameter from the prediabetic period. Toassesstheresponsetoadrenergicstimulation, epinephrine Group3 Group4 was infused ata rate of0.05 and0.1 og/kgpermin. Theextra- Control 12mo Control 12mo systolic threshold was reduced to a significantly greater degree in the nontrained diabetics vs. group 1 controls (Fig. 2). In the mA mA mA mA traineddiabetics, ontheotherhand, thethreshold atthehigher A. Repetitiveventricularextrasysoles dose was significantly greater than in the nontrained diabetic, didnotdifferfromthecontrolsofgroup 1,butwassignificantly 35 25 37 36 lessthanthetrained normals. 47 22 41 45 DuringepinephrineinfusionmeanplasmaK+levelsdeclined 43 27 46 44 10%fromtheinitialnormallevelsbuttherewasnodifference 36 28 _ - Mean±SE 40.3±2.8 25.5±1.3t 41.3±2.6 41.7±2.8 70 CONTROLS DIABETICS B.Fibrillationthreshold 60 _ P-0.001 Figure1. Thresholdsforrepeti- 50 tiveventricularectopicrespon- 48 22 41 37 X 40 ses.o,exercised.Group3,the 42 20 43 46 nonexerciseddiabetics(n = 9), 47 31 49 48 - 30 hadasignificantreductionof 38 22 20 _ 0 0 threshold comparedwithgroup Mean±SE 43.8±2.3 23.7±2.4t 44.3±2.4 43.7±3.4 1 controls(n = 11). Inthe o0 traineddiabeticsofgroup4(n GRP= 9),thethresholdwassignifi- * Individualanimalsthatwerestudiedbeforeinductionofdiabetes GROUP GROUP GROUP GROUP cantly higherthaningroup3 and 1 yrlater. 102P-0.032 4X diabetics, lowerthangroup2(n tPairedttest,P<0.025 (one-tail): P<0.05 (two-tail). N.S. = 5),andsimilartogroup 1. §Pairedttest,P<0.0025 (one-tail);P<0.005(two-tail). 386 Bakthetal. 60 withtrichromewerenegativeingroups 1 and2(Fig. 3A). How- _-a ,- TRAINED ever,inthesedentarydiabeticsofgroup3,accumulation ofblue- 50 - 1I stained material waspresentaround intramural vesselsandac- cretionswere presentthroughout interstitium (Fig. 3 B), which en 40 wasnegligibleingroup4. Onelectron micrography, ultrastruc- a. ture ofcell organelles in the sedentary diabetic appeared to be 30 . - P<O.003 within normal limits(Fig. 4). Theinterstitium containedfibrils j withtheperiodicityofcollagenandamorphousmaterialaround 20 I the capillary and nerve endings, as well as between adjoining cardiaccells. BASAL 0.05 0.1 Left and right ventricularglycogen concentrations were el- EPINEPHRINE (,ug/Kg/min) evatedingroups2and4butnotinthesedentarydiabetic(Table Figure2. Repetitive ventricularresponseafterepinephrine. Thenon- VIII). Ingroups3 and4,cholesterolandtriglycerideconcentra- exercisedgroup3diabetics(o)hadalowerthresholdatthehigherepi- tions were increased in both ventricles. Phospholipid concen- nephrinedosethantheexerciseddiabeticsofgroup4(A)andgroup 1 trations in left ventricle for groups 1 through 4, respectively, controls(o).Group2wassignificantly higherthangroup4. were 16.8±0.7MM/gm, 17.2±0.9, 16.7±0.8,and 17.0+0.6.There wasalsonosignificantdifferenceinrightventricularconcentra- II ofthe EKG times did not differ in the diabetic groups vs. tionsbetween thesegroups. Sodium wasalsoincreased butpo- control(TableVI). Intheexercisedcontrolsofgroup2,P-Rwas tassium concentrations were normal in both diabetic groups significantly longer. The H-Q time was significantly prolonged (TableVIII). Watercontentaspercentofleftventricularweight indiabeticsofgroup3comparedwiththenonexercisedcontrols did notdiffersignificantly between the fourgroups. Ingroup 1 ofgroup 1.ExerciseddiabeticshadH-Qtimesthatdidnotdiffer the value was 78.1±0.06%; group 2, 78.3+0.04; group 3, from theexercised controlsofgroup 2. 79.8±0.07; group4, 78.5±0.05. Myocardialcomposition. Theconcentrations ofcollagen in Myocardialischemia. Toassesstheincidenceofspontaneous theleftand rightventriclesaswell asseptum were significantly ventriculararrhythmias, we prepared additional animalstode- increased ingroup 3 diabetic animals(Table VII). These incre- terminetheresponsetoa 15-minperiodofmyocardialischemia. mentsabovecontrolswereproportionatelysimilarineachtissue, Thesedentary diabetic animals hadaglucose clearance rate of although concentrations were highest in the right ventricle of 2.05±0.19, comparable tothelevelsingroup 3 diabetics. After controlsanddiabetics. Collagen incrementsintheleftventricle anestheticinduction,heartrateandaorticpressurewereatsimilar were moreevident inthesubendocardium butthiswasnotsig- levels in the control and diabetic animals before the onset of nificantly greater than in the outer layers. Similar increments ischemia, andplasmaelectrolyteswerenormal. Duringcoronary werenotobservedintheexerciseddiabeticsofgroup4inwhich occlusion the aorta to peripheral coronary artery pressure gra- collagenconcentrationsapproximate thenormalcontrollevels. dientwas 88±4.9 mmHginthecontrolsand91±6.4 mmHgin Toassesstheeffects ofthe hypertrophy ofexercise on collagen thediabetics,andthesinusrateresponsewascomparable(Table accumulation, thecontentpertotalventriclewascalculatedfrom IX).Thenumberofectopicbeatsasventriculartachycardiawas muscleweightandcollagenconcentration (TableVII). Content significantly greater in the diabetic animals (Table IX). Fibril- was significantly greater in both the left and right ventricles as lation occurred in fouroftheseven diabeticsand intwo outof wellasseptumingroup3butgroup4didnotdifferfromcontrols. 12 inthenormalgroup. Whenthosewith fibrillation andthose Morphologic studies were performed to determine the dis- with ventricular tachycardia alone were compared with nondi- tributionofcollagen. Histologicsectionsofleftventriclestained abetics, the difference was significant at P < 0.05. The area of Table V ResponsestoEpinephrineInfusion inDiabetics Left Aortic ventricular Freefatty pressure end-diastolic K+ acid Heartrate mean pressure Vces mEqlliter mEqlliter beats/min mmHg mmHg Group 1 Basal 3.9±0.16 776±86 122±7 138±8 5.1±0.6 6.8±0.7 Epinephrine(0.1 tg/kgpermin) 3.6±0.16 1,478±234 127±9 126±10 3.9±1.3 10.6±1.9 Group3 Basal 4.1±0.12 716±96 133±10 140±4 4.9±0.7 7.5±0.5 Epinephrine(0.1 tg/kgpermin) 3.9±0.32 1,331±232 145±7 136±9 2.6±0.5 9.9±1.2 Group4 Basal 4.1±0.17 762±82 110±12 138±9 3.4±0.7 7.7±0.4 Epinephrine(0.1 sg/kgpermin) 3.7±0.22 1,460±310 117±15 124±10 3.7±1.2 12.0±1.4 Unpairedttest. Nosignificantdifferencesinresponsetoepinephrinebetweengroup 1 vs.groups3and4. *Ratio ofmaximaldP/dttosimulta- neousleftventricularpressureincyclelengths/s. Diabetes, CardiacComposition, and Vulnerability 387 Table VI. Myocardial Conduction inNormals andDiabetics ening at ADP50 and APD80 after isoproterenol in the diabetic group, whichdid notoccurinnormals. Heart rate PR QRS QTc HQ Discussion beats/ min ms ms ms ms Theseexperimentswereundertakentodeterminewhetheramild Group 1 form ofdiabetes, 1 yrinduration,wouldalterthesusceptibility (n= 6) 128±11 71±7 65±2.6 324±13.7 25.0±0.8 ofthe myocardium toarrhythmias. Thiscanine model ischar- Group2 acterized by normal coronary arteries (2) and collateral blood (n = 5) 94±15 96±8* 68±1.1 310±6.5 28.5±0.7f flowresponseduringcoronaryocclusion(34),aswellaspreserved Group3 ultrastructure ofcardiac cells(8). (n = 6) 133±10 67±6 63±0.78 312±7 29.7±0.98t Sedentarydiabetics. Diabetic dogsthatwere notphysically Group4 conditioned showedchangesinelectrical vulnerability, withre- (n = 6) 100±12 84±6 64±3.3 307±10 31.6±1.6t duction ofboth the thresholds for repetitive extrasystoles and ventricular fibrillation. Moreover, agreater incidence ofspon- Unpairedttestvs.Group7. *P<0.05. *P<0.01. taneousventriculararrhythmiaswasevidentduringacuteisch- emia ofthe anterior left ventricle. That arrhythmogenesis was affected in both ventricles with similar compositional abnor- risk intheregion ofthe anteriordescending artery was 33±1.8 malities issupported bythisobservation. ginthe normal vs. 31.7±2.1 ginthediabeticgroup. Reduction ofvulnerability thresholds in the sedentary dia- Actionpotentials. Todetermine whether electrophysiologic beticsmayberelatedtoseveral factors. Connectivetissuesepta abnormalities were present in the sedentary diabetics, group normally presentwithin individual musclebundlesinaninter- 3-A was compared with group 1-A normals (Table X). In the mittentdistributionareconsideredtobeassociatedwithlocalized basal state there was no difference in the MDP orAPA in the dissociation ofexcitation (39).Accordingly, theincreaseofcol- myocardial cell between normals and diabetics. However, the lagencontentinthediabeticmyocardiuminanonhomogeneous duration ofrepolarization at 50 and 80% was slightly, but sig- distributionmayexaggeratethisphenomenon. Thisviewissup- nificantly, longer inthediabetic group(P<0.02). ported by the local electrode study, which revealed a modest After isoproterenol, normals exhibited a significant short- increaseofHis-Qtime,approximatingthatofanethanolmodel eningoftheAPD50orAPD80withoutachangeinMDPorAPA of subclinical heart disesase (22). Assuming that this at least (Table X). This response was significantly greater in diabetic partially reflects subendocardial collagen accumulation, alocal animals.Aplotofthelogconcentrationoftheadrenergicagonist dispersionofrefractoryperiodswouldfacilitatereentrantactivity vs. the action potential duration in myocardial cells reveals andarrhythmias(40).Thismechanismhasbeenconsideredthe greater shortening at both repolarization durations (Fig. 5). In basis forarrhythmias duringthreshold stimulation (41) aswell thepresence ofsimilarslopes the regression line levels differed asearly in acute ischemia (42) in nondiabetics. Further, under indiabetics vs. normals (P<0.05). appropriate conditions, triggered activity maybeelicited inthe In the Purkinje fibers ofthe diabetics the MDP, APA, and diabetic (43). therepolarization durationswerenotsignificantly differentfrom Catecholamine responses. Several observations supportthe normalinthebasalstate.Therewas,however, significantshort- view that catecholamines may influence arrhythmia suscepti- bilityindiabetics. Exogenousepinephrinewasfoundtoreduce thethreshold forrepetitive ectopic beats in sedentary diabetics Table VII. OH-ProlineConcentrations* andto shortentherepolarization phases oftheactionpotential relativetonormals. Duringthevulnerabilitytestthesympathetic Leftventricle system inthemyocardium appearstobeactivatedatthesiteof Group Inner1/2 Outer1/2 Septum Rightventricle thestimulatingelectrode, evenbythesingle-impulse technique (16).Thus,anenhancedsensitivitytocatecholaminesindiabetics 1 (n = 11) 2.15±0.11 2.09±0.15 1.72±0.09 2.88±0.10 may partially account for the reduced fibrillation threshold in 2(n = 5) 2.16±0.28 1.88±0.26 1.64±0.21 3.0±0.29 thebasal state. 3(n = 9) 3.14±0.24t 2.63±0.27§ 2.41±0.28§ 3.56±0.27§ During ischemia a neurohormonal contribution to arrhyth- 4(n = 9) 2.11±0.17 2.02±0.14 1.83±0.22 3.08±0.09 mogenesisisattestedbytheprevention oflethal ventricularar- rhythmias after beta-blockade or chemical sympathectomy in OH-ProlineContent, MilligramsperRegionofHeart nondiabetics (44). While a similar antiarrhythmic efficacy in Group Leftventricle" Septum Rightventricle mild diabeteswouldsuggest adominant role forthis system, a quantitativecomparison with normalswouldrequireassessment 1 40.9±2.9 15.4±0.7 23.3±1.2 ofthe dose-response relationship to beta-adrenergic blockade. 2 42.4±5.9 16.0±2.5 26.7±3.4 Althoughareductionofvagalactivitycancontributetoarrhyth- 3 59.8±3.3t 25.0±2.2t 30.9±3.0§ mogenesis (45, 46), in experimental diabetes the efferent limb 4 40.7±5.0 17.6±2.4 25.6±1.7 ofthis nerve has been reported to be unimpaired asjudged by the heart rateresponseto vagal stimulation (47). *Concentrations in ug/mgdryweight. tP<0.01. §P<0.02. Altered responsiveness to catecholamines appearsto be se- Contentsininnerandouterhalvesofmyocardiumwereaveragedfor lectivesincethemyocardialhemodynamicresponsedidnotdiffer eachanimal. between control and diabetic animals. In adiabetic rat model, 388 Bakthetal 3 e~~~~~~~ }~~~~~~~~~~~~~ >c 8 0 8UA 0 6~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I.= Jo 0 WI-c t. I'd~~~~~~~~~~~~~~~~~~~~~C S_ _~~~~~~~~~~~~~~C O Tu. #E .tX~~~~~~~~~~~~~~~C V.~ ~ ~ III fv Diabetes, CardiacComposition, andVulnerability 389 r. 1! I'S .r-. ..: K. 0 H. ti~ A... ~~~~~~~~~~~~ iSL i ;ip~. C % RI .1 M' A E Fi.urnsrisinelcro icorahfrma eenay ibei o rop3 X1150 eor edcio) Acmuaio f irilr n amoapthisurhsieroagnhucuotfth inertitumpariclaly n roxmix t te cpdarxan nevefibrs Soe fte ibrlsar w .1 390Baktahal.~~~~~~~~~~~~~~prl 00 (NI bX +1 +1 +1 +1 z: z :4 N N i N (NI (4 C4 2 - N 4t +1 +1 +1 +1 0. z z P. N 9 No N 00 (N - (N 0+1 -+1 ;I- 6 o~ C Vv , 14 ~o +1 +1 w _ ^~~w _~ N ( - 2 +1 +1 +1 +1 o. oo.v cv)GOV v sq D. J ON (IC z~z 6 6 6 6 'F +1 +1 +1H ++11 0N 0 Z CA4 4 Z -V(A 0o t +1 +1 N+l 0+%1 C-, I _en ~. f 99 6 6 o_.C. '0 '.. CO ~o 9_~ i -) N 4 i 6o E _ _'I 6+l 6+l Zn 4 4 Z1 en e Q C- Ca.qZ III>tN +N1 0+01 6 6 :v C-V4v~~~~43 .0 11 u 900 * (NI 'C0q~ b ~+1 +1 +1 +1 Z b O0 0 ~ 'I 00 ',o C4 cn en C14~~~~~~4 U akm5-00-0''C:;-, i0. bkE- r+1O +1(eQ0%+1^ 0%+1 VI4')a-00V~i~~QCoZoqO~~V)40.o0VQ0~oV~4V~o~~,w8~0~~~4 vvut .I .SII . 0 ~~"i~~~~w4) 2421+ 1 + u6 oQ oN6o =o~ on ononin Diabetes, CardiacComposition, and Vulnerability 391
Description: