OnPumpIncompletenessRevascularisationCardiovascularUniversityMonashHe - PDF document

On-PumpIncompletenessRevascularisationCardiovascularUniversityMonashHeart,Health,Melbourne,DepartmentSciencesUniversity,acceptedpublished-ahead-of-print SocietySurgeons(ANZSCTS)ZealandElsevierrights *CorrespondingauthorMonashCardiovascularResearchCentre,UniversityMonashHeart,Health,adam.brown@monash.edu BackgroundOff-pump coronary artery bypass grafting (CABG) negates the requirement for extracorporeal circulationused with the traditional on-pump approach. However, off-pump CABG is technically more challengingand may theoretically lead to less complete revascularisation. Recent data suggests a prognostic benefit fortraditional on-pump CABG, but the mechanism for this remains unclear. We hypothesised that the inferioroutcomes with off-pump CABG could be driven by the need for repeat revascularisation, with this benefitonly becoming clear at long-term follow-up. We therefore evaluated short, medium and long-term out-comes of patients undergoing revascularisation with on vs. off-pump CABG.Electronic databases were searched to identify suitable randomised controlled trials enrolling
100 patientsin each arm. Clinical outcomes were extracted at 30-days, 12-months or �4 years. The primary outcome waslong-term all-cause death, while secondary outcomes included 30-day, 12-month and �4-year cardiacdeath, stroke, myocardial infarction or revascularisation.Thirteen (13) studies comprising 13,234 patients were included. Off-pump CABG was associated with anincreased risk of all-cause death (Odds Ratio [OR] 1.18, 95% confidence interval [CI] 1.02–1.32, p = 0.01) andrepeat bypass surgery (OR 2.57, 95%CI 1.23–5.39, p = 0.01) at long-term follow-up. A significant, increasedrequirement for revascularisation in off-pump was seen at 12-month follow-up (OR 1.59, 95%CI 1.09–2.33,p = 0.02). No differences were noted between groups at 12-months and �4 years for myocardialinfarction or stroke.Off-pump CABG is associated with significantly higher rates of all-cause mortality rate at long-term follow-up. These outcomes demonstrate a temporal relationship that may be driven by a greater requirement forrepeat revascularisation at 1- and 5-year follow-up.registration (PROSPERO).KeywordsCoronary artery by

pass graftingOn-pumpOff-pumpMortalityMeta-analysis 1443-9506/04/$36.00https://doi.org/10.1016/j.hlc.2018.11.019 Introductionction. Furthermore, the extracorporealcircuit in cardiopulmonary bypass may lead to additionaldeleterious effects such as cardiac ischaemia, neurocognitivedysfunction, coagulopathy necessitating requirement forblood products, atrial brillation and the systemic inamma-tory response syndrome [2,3].Coronaryarterybypassgraftingwithouttheuseextra-corporealcircuit,theso-called‘off-pump’approach,wasestab-lishedovercometheselimitations.Off-pumphasbeenassociatedwithreducedperioperativemorbidityincludingreducedstrokerates,reducedlengthstay,reducedbloodproductrequirementandreducedrenaldysfunction[4,5]Despitetheseimmediatebenets,thereremainsongoingcon-cernsurroundingthelong-termdurabilitytheoff-pumpapproach,completerevascularisationusingthistechniquecanbedifculttoachieve.Off-pumpCABGisalsoconsideredatechnicallymorechallengingprocedurethatrequiressigni-cantoperatorexperience[6]Randomisedtrialscomparingon-versusoff-pumpCABGconicting[7,8]whilepreviousmeta-analysesincludingobservationalstudiesreportcontrast-ingoutcomesbenetrisk[5,9–12]Themostrecentpooledanalysisrandomisedtrialslong-termfollow-updemonstratedstatisticalsuperioritytheon-pumpapproach[13]butprovidedclearmechanismforthesurvivalbenetobserved.Thisparticularstudyalsosignicantlyham-peredstatisticalmodelthatdidappropriatelyaccountforvariabilitybetweenclinicaltrialsandconfoundedbias[14–16]Thismayhavemajorimplicationsthegreaterthevariancebetweentrials,thegreaterimpactthenalpooledestimate.Althoughstudieswererandomisedcon-trolledtrials,comparativepopulationcohortswerenotablyheterogeneous.MethodsDatawerespeciedpriorconduct-literaturesearch.Itemsfordataoutcomesmor-tality,stroke,myocardialrevascularisation,atrialbrillationoccurrence,anddysfunction.(N.N.U.T.)dentlyconductedsearchandperformedstudiesextractedwereveried(A.J.B)withresolvedconsensus.withinstudieswereevaluatedNewcastle-Ottawascaleale.Clinical Endpoints U. Thakur et al. . Publication bias was visually assessed byfunnel plots and statistically by the Egger test. A two-sided pvalue of 0.05 was considered signicant.ResultsA total of 2,524 potential citations were sc

reened, with 237studies identied for potential inclusion and further evalua-tion. Of these, 224 studies were excluded as they did notcompare on-pump with off-pump CABG (n = 133), therewere 100 patients in either study arm (n = 69), they weresubset analyses of a different RCT (n = 2), follow-up time wasunclear (n = 3) or the study reported a different outcome ofone of the included trials (n = 17). Full identication andprocess of study exclusion are detailed in the PRISMA owdiagram (Supplemental Figure S1).Thirteen studies with a total of 13,234 patients met the pre-specied inclusion criteria and were included in the nalquantitative analysis [7,8,18–28]. Some studies comprised thesame cohort but were published separately for the differenttime points of 12-month and 30-day outcome [29–34]. Furtherdetails of study patient demographics are presented in Sup-plemental Tables S5–S6.Long-Term OutcomesTherevereportedmortalityfollow-up[7,8,18,21,28]Theresignicantincreasetheriskall-causepatientsundergoingoff-pumpCABG:1.03–1.34,0.01)FigureThereweredifferencesgroupsthe(OR95%0.56–1.11,diacdeath1.07,95%0.90–1.26,myocardialinfarction1.13,95%0.79–1.61,0.52)revascu-larisation0.95–1.41,0.52).However,thereincreasedrepeatoff-pumpCABGgroup2.57,1.23–5.39,Sensitivityanalysisusinghazardestimatesforall-causemortalityavail-fromthreestudies[7,8,18]demonstratedincreasedeathoff-pumpgroup(pooled95%1.02–1.30,StudyreportedeventandsummarypooledestimatesreportedTablesandrespectively. 1Forest plot of the primary outcome of long-term all-cause mortality. Forest plot demonstrates an increased risk ofall-cause mortality in patients undergoing off-pump coronary artery bypass grafting, odds ratio 1.18, 95% condenceinterval 1.03–1.34, p = 0.01. vs. Off-Pump CABG 151 Outcomesmes. With regards to the peri-procedural phase, we found no signicant differencebetween treatment arms in the incident of all-cause death(OR 0.99, 95% CI 0.77–1.26, p = 0.91), stroke (OR 0.82, 95% CI0.61–1.10, p = 0.18), myocardial infarction (OR 1.00, 95% CI0.73–1.37, p = 0.99) or revascularisation (OR 2.29, 95% CI0.76–6.90, p = 0.14). Additionally, we found no differencein the incidence of renal failure (OR 0.89, 95% CI 0.65–1.22, p = 0.46) or postoperati

ve bleeding (OR 1.12, 95% CI0.60–2.09, p = 0.72) between surgical strategies. There was aborderline signicant increased risk of atrial brillation infavour of off-pump CABG (OR 0.57, 95% CI 0.32–1.00,p = 0.05).BiasThere was no evidence of small study effects bias (p = 0.50) orbias by visual assessment of funnel plots for the primaryoutcome. Using the bias review tool, all studies contributingto the primary endpoint demonstrated only low risk of biasaside from the category of blinding with respect to the type ofsurgical technique (Supplemental Table S7).Discussion All-CauseRevascularisation,Follow-Up-Up 14/12 NR NR NR NRCORONARYARY 15/14 2.3/2.7 10/9.7 7.4/8.1 2.8/2.3MASS III [21] 8.4/5.2 1.9/3.3 NR 6.5/2.0 6.5/5.9OCTOPUSS 8.5/6.5 1.4/3.6 0/1.4 4.9/6.4 7.7/5.0ROOBY [7] 15/12 NR 6.3/5.3 12/9.6 13/1212-MonthCORONARYARY 5.1/5 1.5/1.7 4.2/4 6.8/7.5 1.4/0.8OCTOPUSS 1.4/1.4 0.7/1.4 NR NR NRROOBY [33] 3.9/2.7 NR 2.6/1.3 NR NRGOPCABEBE 7/8 3.5/4.4 NR 2.1/2.4 3.1/2.0Mazzei et al. [23] 3.3/2.7 NR NR NR NR30-Day Follow-UpBHACAS 1 [29] 0/2 NR NR 1/4 NRBHACAS 2 [29] 0/0 NR NR 0/1 NRCORONARYARY 2.5/2.5 1/1.1 NR 6.7/7.2 0.7/0.2OCTOPUSS 0/0 0.7/1.4 NR 4.9/4.3 NRROOBY [33] 1.6/1.2 1.3/0.7 NR NR NRDOORS [20] 1.6/1.8 2.2/4.0 NR 8.2/5.6 NRGOPCABEBE 2.6/2.8 2.2/2.7 NR 1.5/1.7 1.3/0.4BBS [24] 3.4/6.7 4.0/3.7 NR 5.1/9.2 0.6/1.8PRAGUE 4 [26] 2.0/1.1 0/1.1 NR 2.0/1.6 NRMotallebzadehzadeh 1.9/1.0 0.9/2.9 NR NR NRMazzei et al. [23] 2/1.3 NR NR NR NRIqbal et al. [27] 2/3 2/6 NR 11/2 NRLemma et al. [22] NR 0/0.5 NR 1.9/3.0 NRTableoff-pump/on-pumppercentagesAbbreviations:High-RiskPatientsRandomizedOff-PumpCoronarySurgeryoff-pumpsurvival,long-termfollow-uprandomizedCORONARY,Off-PumpCoronary-ArteryCORONARYOutcomesCoronary-ArteryGrafting;SurgeryElderlyPatientsDanishVersusOff-PumpRandomizationGOPCABE,versusCoronary-ArteryElderlyPatients;Five-YearRandomizedOff-PumpGrafting;infarction;reported;OCTOPUS,CardiacOff-PumpCoronaryGraftSurgery;RandomizedComparisonBypassExtracorporealCirculationOff-PumpCoronarySurgery;Five-YearOutcomesOff-PumpCoronary-Artery U. Thakur et al. 5-year clinical follow-up (Figure 2) and may be driven by ahigher rate of repeat CABG in off-pump CABG patients. Nodifferences were seen in rate of stroke, myocardia

l infarction(MI) or cardiac death at any time-point between surgicalstrategies. One-year outcomes similarly demonstrated anincreased risk of repeat revascularisation in off-pump CABGpatients however, at this stage, there was no associatedmortality difference. Finally, there was a borderlineincreased risk of postoperative atrial brillation (AF) favour-ing off-pump CABG patients with no difference in peri-procedural death, stroke, MI, or revascularisation.Inconsistent conclusions have been raised from previousmeta-analyses of this topic, likely due to inclusion of largeamounts of observational data, disparate composite end-points, small sample sizes and short clinical follow-up dura-tion. Smart et al. recently reported a similar nding of long-term benet with on-pump CABG without a clear underlyingmechanism as well as using a potentially awed methodol-ogy as reported by several letters of correspondence since thepublication of this study[13–16]. Our analysis benets fromthe inclusion of only large-sample, high-quality RCT datawith comparable individual endpoints, a statistical method-ology to account for between and within trial heterogeneity,and a time-related analysis to assess the evolution of out-comes. We believe these criteria were required to establishthe necessary power and duration to detect meaningful dif-ferences between groups.A temporal relationship for the benet of surgical inter-vention has been previously described with CABG in com-parison to percutaneous coronary intervention (PCI) andmedical therapy[35–37]. Similarly, our data demonstratean improvement in mortality over time in on-pump CABGpatients, which was not observed at 30-day and 12-monthfollow-up. This delayed effect may be driven by less com-plete revascularisation at surgery in the off-pump CABGpatients. Given the technical challenges involved in off-pump CABG with surgery on a beating heart especially whensignicantly enlarged, fewer bypass grafts and distal anas-tomoses may be performed and our results noted anincreased rate of any revascularisation at 12-months, andincreased requirement for repeat CABG at �4 years in the Follow-UpMortalityRevascularisation12-MonthMortalityRevascularisationMortalityRevascula

risationReoperationAbbreviations:atrialarteryconfidenceinterval;myocar-dialfavouringOFF-PUMP);percutaneousintervention. 2Temporal trends in mortality. Graph demon-strates pooled odds ratio (solid circle) with respective95% condence interval (dashed line) for all-cause mor-tality at 30-day, 1-year and 5-year follow-up. Sold hor-izontal red line is the line of unit (odds ratio = 1.00). Thisdemonstrates no signicant difference between groupsat 30-day and 1-year follow-up, however a statisticallysignicant association with all-cause mortality at the 5-year time point in favour of on-pump suggesting atemporal association with the outcome. vs. Off-Pump CABG 153 ConclusionsAcknowledgementsSourcesDisclosuresAppendixSupplementaryReferencess Nerlekar N, Ha FJ, Verma KP, Bennett MR, CameronPercutaneous U. Thakur et al. stenosis:meta-analysis2016;9(12).2). Weiland AP, Walker WE. PhysiologiccardiopulmonaryLung1986;15(1):34–9.15(1):34–9. Fudulu D, Benedetto U, Pecchinendaon-pumpcoronarycontrolled2016;8(Suppl10):S758–71.S758–71. Kowalewski M, Pawliszak W, MalvindiBokszanskiPerlinskigraftinghigh-riskpatientscomparedon-pumpMeta-analysis.Cardiovasc2016;151(1):60–77.7. Deppe AC, Arbash W, Kuhn EW, Slottosch I, Scherner M, LiakopoulosarteryinvestigatedrandomizedEurCardiothorac2016;49(4):1031–41.discussionsion Puskas JD, Mack MJ, Smithoff-pump2010;362(9):851.853–854.. Shroyer AL, Hattler B, Wagner TH, Collins JF, Baltz JH, Quin JA, et al.Five-year outcomes after on-pumpcoronary-artery2017;377(7):623–32.32. Lamy A, DevereauxPrabhakarancoronary-artery2016;375(24):2359–68.. Filardo G, Hamman BL, da Graca B, Sass DM, Machala NJ, Ismail S, et al.Efcacy and effectivenessversusmortalitysurvival.Cardiovasc2018;155(1):172–9.e175.5. Takagi H, Ando T, Mitta S, group A. Meta-analysis comparingoff-pumpon-pump2017;120(11):1933–8.1):1933–8. Zhao DF, Edelman JJ, Seco M, Bannon PG, Wilsonmeta-analysis.Cardioliol Moller CH, Penninga L, Wetterslev J, Steinbruchelon-pumpCochraneDatabase):2012;(3)CD007224.007224. SmartNA,KingLong-termoutcomesversusoff-pumpcoronaryarterybypassgrafting.CollCardiol2018;71(9):983–91.. Ma G, Fan Y, Shao W, Qi LW. Meta-analysis2018;72(3):344–5.. Shah R. Long-term survival with on-pumpp Zhao DF, Edelman

JJ, Seco M, Bannon PG, Vallely MP. Long-termfollowingoff-pumpxed-effectsrandom-effectsmodels.Cardiol2018;72(3):345–7.(3):345–7. Stang A. Criticalevaluationnewcastle-ottawanonrandomizedstudiesmeta-analyses.2010;25(9):603–5.(9):603–5. Angelini GD, Culliford L, Smith DK, Hamilton MC, Murphy GJ, AscioneEffectsoff-pumpfollow-upcontrolledCardiovasciovasc Diegeler A, BorgermannUrsulescuOff-pumpcoronary-arterypatients.2013;368(13):1189–98.9–98. Houlind K, Kjeldsen BJ, Madsen SN, RasmussenHolmebypasson-pumpoff-pumprandomizationCirculation2012;125(20):2431–9.1–9. Hueb W, Lopes NH, Pereira AC, Hueb AC, Soares PR, Favarato D, et al.Five-year follow-uprandomizedcoronarymass2010;122(11Suppl):S48–52.):S48–52. Lemma MG, Coscioni E, TrittoFondaconerandomizedThoracCardiovasc2012;143(3):625–31.3(3):625–31. Mazzei V, Nasso G, SalamoneTommasiniAnselmirandomizedgraftingextracorporealsystemoff-pumpCirculation2007;116(16):1761–7.7. Moller CH, Perko MJ, Lund JT, Andersen LW, Kelbaekhigh-risk2010;121(4):498–504.):498–504. Motallebzadeh R, Bland JM, Markus HS, Kaski JC, Jahangirifunctioncerebralemboli:randomizedon-pumpThorac2007;83(2):475–82.2. Straka Z, Widimsky P, Jirasek K, Stros P, Votavaprague-4.2004;77(3):789–93.(3):789–93. Iqbal J, Ghaffar A, Shahbazwithoutcardiopulmonary2014;26(2):123–8.2):123–8. van Dijk D, Spoor M, HijmanCognitive2007;297(7):701–8.):701–8. Angelini GD, Taylor FC, ReevesAscionemidtermon-pumpbeatingstudiesanalysistrials.2002;359(9313):1194–9.313):1194–9. Lamy A, DevereauxPrabhakaranOff-pumpcoronary-artery2012;366(16):1489–97.97. Lamy A, DevereauxPrabhakaranon-pumpcoronary-artery2013;368(13):1179–88.9–88. Van Dijk D, Jansen EW, Hijman R, NierichDiephuisMoonsCognitiverandomized2002;287(11):1405–12.12. Shroyer AL, GroverMcDonaldversuscoronary-artery2009;361(19):1827–37.. van Dijk D, NierichDiephuison-pumpcoronaryrandomized2001;104(15):1761–5):1761– Molstad P, Moer R, Rodevandpercutaneousintervention.2016;3(2):e000489.00489. Velazquez EJ, Lee KL, Deja MA, Jain A, Sopko G, Marchenko A, et al.Coronary-arterypatientsventricular2011;364(17):1607–16.7–16. Velazquez EJ, Lee KL, Jones RH, Al-Khalidi HR, Hill JA, Panza JA, et al.Coronary-arterybypasspatientscardiomyopa-2016;374(16):1511–20. vs. Off-Pump CAB