ipc對動(dòng)脈旁路移植術(shù)后血液動(dòng)力學(xué)的影響

1/1ipc對動(dòng)脈旁路移植術(shù)后血液動(dòng)力學(xué)的影響IPC對動(dòng)脈旁路移植術(shù)后血液動(dòng)力學(xué)的影響HaemodynamiceffectofintermittentpneumaticcompressionofthelegafterinfrainguinalarterialbypassgraftingK.T.Dells1,M.J.Husmannl,G.Szendro1,N.S.Peters2,J.H.N.Wolfe1andA.O.Mansfield1RegionalVascularSurgeryUnitand2DepartmentofAcademicCardiology,StMary’sHospital,ImperialCollegeSchoolofMedicine,London,UKCorrespondenceto:Mr.K.T.Dells,4AbingerCourt,34GordonRoad,LondonW52AF,UK(e-mail:k.delis@ic.ac.uk)Background:Intermittentpneumaticcompression(IPC)mayincreasebloodflowthroughinfrainguinalarterialgrafts,andhaspotentialclinicalapplicationasbloodflowvelocityattenuationoftenprecedesgraftfailure.ThepresentstudyexaminedtheimmediateeffectsofIPCappliedtothefoot(IPCfoot),thecalf(IPCfoot)andtobothsimultaneously(IPCfoot+calf)onthehaemodynamicsofinfrainguinalbypassgrafts.Methods:Eighteenfemoropoplitealand18femorodistaiautologousveingraftswerestudied;allhadarestinganklebrachialpressureindexof0.9ormore.Clinicalexamination,graftsurveillanceandmeasurementofgrafthaemodynamicswereconductedatrestandwithin5sofIPCineachmodeusingdupleximaging.Outcomemeasuresincludedpeaksystolic(PSV),mean(MV)andenddiastolic(EDV)velocities,pulsatilityindex(PI)andvolumeflowinthegraft.Results:AllIPCmodessignificantlyenhancedMV,PSV,EDVandvolumeflowinbothgrafttypes;IPCfoot+calfwasthemosteffective.IPCfoot+calfenhancedmedianvolumeflow,MVandPSVinfemoropoplitealgraftsby182,236and49percent,respectively,andattenuatedPIby61percent.Enhancementinfemorodistalgraftswas273,179and53percentrespectively,andPIattenuationwas63percent.Conclusion:IPCwaseffectiveinimprovinginfrainguinalgraftflowvelocity,probablybyreducingperipheralresistance.IPChasthepotentialtoreducetheriskofbypassgraftthrombosis.PresentedtotheInauguralMeetingoftheSocietyofAcademicandResearchSurgery’,London,UK,January2002Paperaccepted9December2003Publishedonline27February.2004inWileyInterScience(pression(IPC)oftheleggeneratesasignificantincreaseinarterialbloodflowinpatientswithperipheralvasculardisease1-3.IPCofthefoot(IPCfoot)increasesarterialcalfinflowby50-90percentinarteriopaths,andbymoreinnormalsubjects1.3.Athreefoldtofourfoldincreaseinthecalfbloodflowhasbeenreportedwithfootandcalfcompression(IPCfoot+calf)inpatientswithsuperficialfemoralarteryocclusion2.4.TheapplicationofIPC.forover4monthsamelioratesintermittentclaudicationandimprovestheanklebrachialpressureindex(ABPI)5.6.Thisisprobablyduetoaflowdependentpromotionofcollateralcirculation5.6.Arterialgraftsarecomplicatedbyearlyocclusioninupto20percentofdistalbypasses7.Failurethatoccurswithinthefirst2or3daysisusuallyduetopoorselectionortechnicalerror;latefailurewithinthefirstmonthresultsfromanumberoffactors,includingtechnicalimperfectionpersistentunderlyingdiseaseandsurfacethrombogenicityofthegraft8.9Ithasbeensuggestedthatthereisacriticalthresholdvelocityrequiredforensuringgraftpatencyintheseearlystages10.11Thelikelihoodofearlythrombosisishighinasmall-calibregraftplacedintoahighresistancecirculation,suchasinthecaseofafemorodistalbypassgraftinthepresenceoflowflowvelocities.Itwashypothesizedthat,analogoustoitsbeneficialeffectonnativearterialbloodflow,theapplicationofIPCmightalsoenhancearterialbloodflowthroughaninftainguinalbypassgraft.Thismighthaveclinicaladvantages,aslowflowvelocityprecedesgraftfailure12.TheaimofthisstudywastocomparetheimmediateeffectsofIPCfoot,IPCfoot+calfandIPCofthecalf(IPCcalf)onarterialbypassgrafthaemodynamicsinpatientswhohadinfrainguinalrevascularizationforischaemia.PatientsandmethodsStudygroupsGraftflowvelocitiesduringIPCweremeasuredin16patientswith18femoropoplitealgrafts(fiveaboveand13belowknee)andin18patientswith18femorodistalgrafts(twoperoneal,nineposteriortibial,fiveanteriortibialandtwodorsalispedis).Therewere13menandthreewomenofmean(s.d.)age65.9(9.6)yearsintheformergroup,and15menandthreewomenaged68.8(7.5)yearsinthelattergroup.Allgraftswereautologousvein.Thefemoropoplitealgraftshadbeeninsertedforamedianof17(interquartilerange7.5-33)monthsandthefemorodistalgraftsforamedianof18(interquartilerange12-26)months.PatientswhohadaninfrainguinalbypasswereassessedbymeasurementofrestingABPIfollowedbygraftsurveillanceusingdupleximaging.Uncompromisedgrafts,definedbyarestingABPIof0-9ormore,andnosignificant(50percentdiameterormore)stenosisondupleximaging,wererecruitedintothestudy.StudydesignThesequenceofIPCapplicationwassubjecttoacrossoverdesign.Toeliminatebiasfromcarry-overeffects,thetwostudygroups,comprising18legseach,hadallsixpossiblecombinationsofIPCsequences(IPCfoot-IPCcalf-IPCfoot+calf,IPCfoot-IPCfoot+calf-IPCcalf,etc.)appliedtotheirlegsandinvestigatedevenly(threelegspersequence).Thesesequenceswereallottednumbers1-18atrandomfromsealedenvelopes,oneforeachnewentry.ExaminationandimagingprotocolGraftbloodflowwasstudiedusingcolourduplexultrasonographywithalinear-array7.5-MHzprobe.Subjectswerescannedinthesittingposition,withtheirlegsdependent,slightlyextendedattheknees,andtheirfeetrestingonalowstool.FootandcalfpadsenablingIPCdeliverywereapplied,andarestingperiodof15minwasallowedforflowstabilization,duringwhichthesubjectremainedseated.Graftbloodflowwasmeasuredfirstatrest,andthenwithIPCactivated.BloodflowwasstudiedonthesixthminuteofIPCaction.Datawereretrievedwithin5sofIPCdeliveryandIPCwasthenswitchedoff.Afterrestingfor10min,bloodflowwasreassessed.ThesecondIPCmodewasthenengagedandflowmeasurementstartedonthesixthminuteofpumpaction.Similarly,bloodflowwasmeasuredbefore,andondeliveryofthethirdIPCmode.Atleastthreereadingswereobtainedandthenthemeanwascalculated.Internaldiameterwasobtainedfromlongitudinalandcross-sectionalgraftviews,inbothsystoleanddiastoleonreal-timeB-modeimaging.GraftdiameterwasmeasuredbothatrestandwithIPC(initial5saftercompression).Atleastthreemeasurementsweremadeateachtimepoint,andthemeanwascalculated.SpectralanalysisofgatedDopplersignalsinsonatingtheentirelumenat60enableddeterminationofmeanvelocity(MV).PositioningofthegateoftheDopplerprobewasmeticulouslymaintainedusingphysicallandmarks.TheMVwasthemeanofeachofthevelocityspectraoccurringduringanintervalofatleastfourcardiaccycles;MVatrestandwithin5sofIPCwascalculatedwithdedicatedsoftwaretracingthewaveformprofiles.Thereproducibilityofthemethodhasbeenreportedpreviously3.VolumeflowwascalculatedbymultiplyingMVbythecross-sectionalareaofthegraft.Datagainedfromcomputer-assistedanalysisoftheprofilesofspectralwaveformsincludepulsatilityindex(PI),andpeaksystolic(PSV)andenddiastolic(EDV)velocities.Dopplerwaveformscontainingaliasing,noiseorwallmotionwerediscarded.Allhaemodynamicevaluationswereobtainedbyinsonatinggraftsatthejunctionofthemiddleanddistalthirdsofthethighbythesameoperator.ImpulseunitAllthreeIPCmodesweredeliveredwithanArtAssist1000unit(ACI,SanDiego,California,USA).Thismechanicalpumpconsistedofapneumaticimpulsegeneratorandtwoinflatableplasticpadsdesignedtofitthefootandcalf.Twolarge-boretubesconnectingtheunitwitheachpadseparatelyofferedtheversatilityofisolating(withclamps)eachpadfromtheother,thusenahlingthreeIPCmodes.Thepumpoperatedatthefollowingpresets:inflationpressure120mmHg,deflationpressure0mmHg,andinflationanddeflationtimes4and16srespectivelywithlPCfoot+calf,theonsetofthefootimpulseprecededthatofthecalfby1s.StatisticalanalysisDataanalysiswasperformedusingnon-parametricstatistics(Minitab8.2;StateCollege,Pennsylvania,USA)andStat-View~4.57(AbacusConcepts,Berkeley,California,USA).IntragroupdatawerecomparedusingtheWilcoxonsignedranktest.IntergroupstatisticswereobtainedusingtheMann-WhitneyUtest.ResultsMeanvelocityRestingMVwassimilarinfemoropoplitealandfemorodistalgrafts(Fig.la).MedianMVincreasedby39percentinfemoropopliteal(P=0.001)and36percentinfemorodistal(P0.001)graftsonIPCfoot,by189percentinfemoropoplitealand170percentinfemorodistalgraftsonIPCcalf(bothP0-001),andby236percentinfemoropoplitealand179percentinfemorodistalgraftsonIPCfoot+calf(bothP0.001).IncreasesinMVweresimilarinfemoropoplitealandfemorodistalgraftsforanyIPCmode.TheMVonIPCfoot+calfwashigherthanthatonIPCcalf(femoropoplitealP=0.019;femorodistalP=0.015)andIPCfoot(bothgroupsP0-001).MVonIPCcalfwashigherthanthatonIPCfoot(bothgroupsP0-001).Fig.1Effectsofintermittentpneumaticcompressionofthefootandcalf(IPCfoot+calf),calf(IPCcalf)andfoot(IPCfoot)onameanvelocityandbpeaksystolicvelocityin18femoropoplitealand18femorodistalgrafts.ValuesareexpressedasmedianandinterquartilerangeDiameterBaselinegraftdiameter,measuredatamiddletodistalthighlevel,wasslightlybutnotsignificantlygreaterinfemoropoplitealthanfemorodistalgrafts(Table1).IPCfoot+calfresultedindiameterincreaseinbothgroups(P0.001andP=0.025respectively).ThediameterincreasewithIPCcalfwasmarkedinfemoropoplitealgrafts(P=0.003).VolumeflowRestingvolumeflowwassimilarinthetwografttypes(Fig.2a).Medianvolumeflowincreasedby76percentinbothfemoropopliteal(P=0-002)andfemorodistal(P0.001)graftsonIPCfoot,by172percentinfemoropoplitealand186percentinfemorodistalgraftsonIPCcalf(bothP0.001),andby182percentinfemoropoplitealand273percent’infemorodistalgraftsonIPCfoot+calf(bothP0-001).IncreasesinvolumeflowweresimilarinfemoropoplitealandfemorodistalgraftsforanyIPCmode.IPCfoot+calfgeneratedahighervolumeflowthanIPCcalf(femoropoplitealP=0.013;femorodistalP=0-006)andIPCfoot(bothgroupsP0.001).VolumeflowwithIPCcalfwashigherthanthatwithIPCfoot(bothgroupsP0.001).PeaksystolicvelocityBaselinePSVwassimilariiifemoropoplitealandfemorodistalgratis(Fig.lb).MedianPSVincreasedby17percentinfemoropopliteal(P=0.13)and18percentinfemorodistal(P-0-002)graftsonIPCfo.t,by62percentinfemoropoplitealand45percentinfemorodistalgraftsonIPCcalf(bothP0.001),andby49percentinfemoropoplitealgraftsand53percentinfemorodistalgratisonIPCfoot+calf(bothP0-001).IncreasesinPSVweresimilarinfemoropoplitealandfemorodistalgraftsforanyIPCmode.PSVon[PCfoot+calfwassinfilartothatonIPCcalf.PSVoneitherIPCfbot+calforIPCcalfwashigherthanthatonlPCfoot(bothP=0.002).EnddiastolicvelocityBaselineEDVwassimilarinfemoropoplitealandfemorodistalgrafts.OnIPCfoot,medianEDVincreasedfrom0atbaselineto2.6cm/sinfemoropoplitealgraftsandfrom0to4-lcm/sinfemorodistalgrafts(bothP0.001).OnIPCcalf,EDVincreasedfrom0to7.2cm/sinfemoropopliteal(P0.001)andfrom0to8.8cm/sinfemorodistalgrafts(Pc0.001).OnIPCfoot+calf,medianEDVincreasedfrom0to8.3cm/sinfemoropoplitealgrafts(P0-001)andfrom0to8.8cm/sinfemorodistalgrafts(P=0.001).IncreasesinEDVweresimilarinfemoropoplitealandfemorodistalgraftsforanyIPCmode.EDMon1PCfoot+calfwassimilartothatonIPGcalf-EDVonIPCcalforIPGfoot+calfwashigherthanthatonIPCfoot(bothP=0.002).Table1EffectofintermittentpneumaticcompressionontheluminaldiameteroffemoropoplitealandfemorodistalgraftsFemoropoplitealFemorodistalDisameter(cm)PDisameter(cm)PAtrest0.565(0.505-0.636)--0.492(0.444-0.557)--IPCfoot+calt0,585{0.520-0-663}0.0010.519(0,448-0.593)0,025IPCcalf0,578(0.521-0.654)0.0030.521(0.443-0.590)0.330IPCfoot0.577(0-511-0.656)0.6400.503(0.433-0.550)0.320Valuesaremedian(interquartilerange).Intermittentpneumaticcompressionwasappliedtothefoot(IPCfoot),calf(IPCcalt)orbothcombined(IPCfoot＋calf).*Versusvalueatrest(Mann-WhimeyUtest).Therewerenosignificantdifferencesbetweenfemoropoplitealandfemorodistalgraftluminaldiametersunderanyconditiontested.Fig.2Effectsofintermittentpneumaticcompressionofthefootandcalf(IPCfoot+calf),calf(IPCcalf)andfoot(IPCfoot)onavolumeflowandbpulsatilityindexin18femoropoplitealand18femorodistalgrafts.ValuesareexpressedasmedianandinterquartilerangePulsatilityindexBaselinePIwassimilarinthetwografttypes(Fig.2b).MedianPIonIPCfootdecreasedby24percentinfemoropopliteal(P=0.040)and21percentinfemorodistalgrafts(P0.001).OnIPCcalf,medianPIdecreasedby56percentinfemoropoplitealgraftsand57percentinfemorodistalgrafts(bothP0.001).OnIPCfoot+calf,medianPIdecreasedby61percentinfemoropoplitealgraftsand63percentinfemorodistalgrafts(bothP0-001).ChangesinPIweresimilarinfemoropoplitealandfemorodistalgraftsforallyIPCmode.PIonIPCfoot+calfwassimilartothatonIPCcalf.PIonIPCcalfandIPCfoot+calfwaslowerthanthatonIPCfoot(bothP=0.004).DiscussionThisstudyhasshownthatIPCenhancesflowthroughinfrainguinalarterialbypassgrafts,irrespectiveofthelevelofdistalanastomosis.FlowwasmaximallyincreasedinbothfemoropoplitealandfemorodistalautologousveingraftswithIPCfoot+calf,followedbyIPCcalf.IPCfootwastheleasteffectivemodeinvestigated,butstillaugmentedflowsignificantly.Volumeflowenhancementwithanyofthe[PCmodeswasassociatedwithasignificantincreaseinmeanflowvelocity.Infemoropoplitealgrafts,volumeflowenhancementwithIPCfoot+calfandIPCcalfwasalsoassociatedwithasignificantincreaseinluminaldiameter.ThiswasnotedinfemorodistalgraftsonlyonapplicationofIPCfoot+calf.TheMVwithin5sofIPCfoot+calfapplicationwas236percenthigherthanbaselineinfemoropoplitealand179percenthigherinfemorodistalgrafts.AllIPCmodesproducedasignificantincreaseinEDV,andanequivalentdecreaseinPI,indicatingamarkeddecreaseinperipheralresistance13.14.Amarkerofimpedanceinthedistalvascularbed,PIincreaseswithvasoconstrictionanddecreaseswithvasodilatation15.Anincreaseinthearteriovenouspressuregradient1,4,16-18isthemainmechanismbehindcalfinflowaugmentationwithIPConsitting.IPCemptiesthedeepveinsand,untilarterialflowrefillsthem,venouspressureremainslowerandthearteriovenouspressuregradienthigher,causingbloodflowtoincrease1,4,16,17.However,thismechanismalonecannotexplainthemagnitudeofflowaugmentationnotedinnormalsubjectsandpatientswithclaudication3.18.Shearstressforcesintheperipheralcirculationuponpressuregradientandflowenhancementmaypromotereleaseofendothelialfactors,suchasnitricoxide1,2,18,whosevasodilatoryactivityonresistancearteriolesshouldaccentuateflowfurther2,I8,19.AnothermechanismenablingIPCtoreducearterialresistanceinvolvestheautoregulatoryreflexes4.Asveinsempty,andformostdeflationtime(16s),venouspressurefallstobelow25mmHg16,17.Thevenoarteriolarandmyogenicreflexesaresuspended,causingperipheralresistancetofall20-23.ThegreaterflowenhancementwithIPCfoot+calfislinkedtoitssuperiorvenoushaemodynaxnicperformance16.Lowflowvelocitiesandgraftsurfacethrombogenicityplaycardinalrolesinthepathophysiologyofinfrainguinalbypassgraftfailure.Bandyketal.12demonstratedthataPSVlowerthan40cm/sinautologousinfrainguinalgrafts,compoundedbylowdiastolicforwardflow,increasedtheriskofgraftthrombosis.Apreoperativeincreaseincoagulationfactors,increasedplateletreactivityandreleaseoftissuethromboplastininbloodenhancethrombogenicity24.Lowgraftflowvelocitiesinthisenviromnentcantriggerplateletactivation,secretionandaggregation,withsubsequentthrombusformationandgraftocclusion25.Afterinfrainguinalgraftinsertion,hyperaemiadevelopsuponperfusion,followedbyanincreaseinoutflowresistance26.EDVdecreasesandPIincreaseswithreactivationofthecapillaryautoregulatorymechanismsinthedistalcirculation.Withincreasingperipheralresistanceinthisphase,flowvelocitiesmaydecreasebelowathromboticthresholdlevel24.Postoperativedifferencesinperipheralresistanceareincreasinglyperceivedtoaffecttheoutcomeofinfrainguinalbypassgrafting26,27.Revascularizationunderthecombinationofgeneralandepiduralanaesthesiaappearstoofferbettergraftpatencyhanthatcarriedoutundergeneralanaesthesiaalone26,27.Thismaybetheresultofperipheralresistancereductionbyepiduralanaesthesia26,27.Inthelightofitshaemodynamiceffects,IPCoffersthepotentialtopreventvelocityattenuationaftersurgery.Thereisonlyalimitednumberofmethodsforenhancingflowininfrainguinalgrafts.Byinducingperipheralvasodilatation,prostaglandinscauseasignificantincreaseingraftbloodflow28.Iloprost,asyntheticanalogueofprostacyclin,generatesashort-lived(20min)decreaseinperipheralresistanceanda52percentincreaseingraftflowwheninfusedintofemorodistalgraftsbeforeperfusion29.Inamulticentrerandomizedtrial,however,graftpatencywasnotimprovedafter1yearoffollowup29.IPChasnonotableside-effects.ThemedianvolumeflowincreasewithIPCrangedfrom76percentwithIPCfoot,forfemoropoplitealandfemorodistalgraftsrespectively,to182and236percentwithIPCfoot+calf.ApplicationofIPChascertainlimitations.Deepveinthrombosisandleginfectionarecontraindications,theflowincreasewanessoonafterIPCisstopped,andeffectiveIPCusepresupposeslimbveinprimingandthuslimbdependency.Furthermore,IPCshouldnotbeapplieddirectlytotheanastomosisandsurgicalwoundsafterbypassgraftingasthismightcompromisetissuehealing.DifferentpneumaticcuffsandmodesincreasetheflexibilityofIPC:IPCfoot+calfcanbeusedinfemoropoplitealgrafts,IPCfootinfemorocruralgrafts,andIPCcalfwithnarrowcuffsinfemoropoplitealorfemoropedalgrafts.Whenthesaphenousveinisharvested,longskinbridgesimprovetheapplicabilityofIPC.Providedthatearlyambulationisnotundulydeferred,IPCmaybeusedforintervalsequaltothoseofsequentialcompressionimplementedfordeepveinthrombosisprophylaxis.AcknowledgementsTheauthorsacknowledgeACI-Medical~,SanDiego,California,USA,fortheloanoffiveArt-Assist1000~unitsthatenabledinvestigationofthehaemodynamiceffectsofintermittentpneumaticcompression.References1MorganRH,CarolanG,PsailaJV,GardnerAMN,FoxRH,WoodcoekJP.Arterialflowenhancementbyimpulsecompression.VascSurg1991;25:8-16.2EzeAR,ComerotaAJ,CisekPL,HollandBS,KerrRP.VeeramasuneriRetal.Intermittentcalfandfootcompressionincreaseslowerextremitybloodflow.AmFSurg1996;172:130-135.3DellsKT,LabropoulosN,NicolaidesAN,GlenvilleB,StansbyG.Effectofintermittentpneumaticfootcompressiononpoplitealarterialhaemodynamics.EurFVascEndovascSung2000;19:270-277.4DellsKT,NicolaidesAN,LabropoulosN,StansbyG.Theacuteeffectsofintermittentpneumaticfootversuscalfvery’ussimultaneousfootandcalfcompressiononpoplitealarteryhemodynamics:acomparativestudy..7FastStag’2000;32:284-292.5DellsKT,NicolaidesAN,WolfeJH,StansbyG.hnprovingwalkingabilityandanklebrachialpressureinsymptomaticperipheralvasculardiseasewithintermittentpneumaticfootcompression:aprospectivecontrolledstudywithone-yearfollow-up.JVascSurg2000;31:650-661.6DelisKT.Arandomisedcontrolledstudyontheeffectofintermittentpneumaticcompressionofthefootandthecalfinthemanagementofintermittentclaudication.InPbDThesis,UniversityofLondon:London,1999;379-405.7RutherfordRB,JonesDN,BergentzSE,BergqvistD,KarmodyAM,DardikHetal.Theefficacyofdextran40inpreventingearlypostoperativethrombosisfollowingdifficultlowerextremitybypass.FVaseSurg1984;1:765-773.8RutherfordRB,JonesDN,BergentzSE,BergqvistD,ComerotaAJ,DardikHetal.Factorsaffectingthepatencyofinfrainguinalbypass.JVascSurg1988;8:236-246.9CheshireNL,TaylorPR,WolfeJHN.Managementoflatecomplicationsfollowinginfrainguinalarterialreconstructions.InLowerLimbIschaemia,MyersKA,NicolaidesAN,SumnerDS(eds).MedOrion:London,1997;497-517.10SauvageLR,BergerKE,MansfieldPB,WoodSJ,SmithJC,OvertonJB.Futuredirectionsinthedevelopmentofarterialprosthesesforsmallandmediumcaliberarteries.SurgClinNorthAm1974;54:213-228.11SauvageLR,WalkerMW,BergerK,RobelSB,LischkoMM,YatesSGetal.Currentarterialprostheses.Experimentalevaluationbyimplantationinthecarotidandcircumflexcoronaryarteriesofthedog.ArchSurg1979;114:687-691.12BandykDF,CatoRF,TowneJB.Alowflowvelocitypredictsfailureoffemoropoplitealandfemorotibialbypassgrafts.Surgery1985;98:799-809.13BurnsP.PrinciplesofdeepDopplerultrasonography.InVascularDiagnosis,BernsteinE(ed.).Mosby:Missouri,USA,1993;249-268.14NichollsSC,KohlerTR,MartinRL,NeffR,PhillipsDJ,StrandnessDEJr.Diastolicflowasapredictorofarterialstenosis.JVascSurg1986;3:498-501.15JohstonKW.ProcessingcontinuouswaveDopplersignalsandanalysisofperipheralarterialwaveforms:problemsandsolutions.InVascularDiagnosis,BernsteinE(ed.).Mosby:Missouri,USA,1993;149-159.16DelisKT,AziziZA,StevensRJ,WolfeJH,NicolaidesAN.Optimumintermittentpneumaticcompressionstimulusforlo