ERCP并发症.pdf

GUIDELINEComplications of ERCPThis is one of a series of position statements discussingthe use of GI endoscopy in common clinical situations.The Standards of Practice Committee of the AmericanSociety for Gastrointestinal Endoscopy prepared this text.This document is an update of a previous ASGE publica-tion.1In preparing this document,a search of the medicalliterature was performed using PubMed.Additional refer-ences were obtained from the bibliographies of the identi-fied articles and from recommendations of expert consul-tants.When limited or no data exist from well-designedprospective trials,emphasis is given to results from largeseries and reports from recognized experts.Position state-ments are based on a critical review of the available dataand expert consensus at the time that the document wasdrafted.Further controlled clinical studies may be neededto clarify aspects of this document,which may be revisedas necessary to account for changes in technology,newdata,or other aspects of clinical practice.This document is intended to be an educational deviceto provide information that may assist endoscopists inproviding care to patients.This position statement is not arule and should not be construed as establishing a legalstandard of care or as encouraging,advocating,requir-ing,or discouraging any particular treatment.Clinicaldecisions in any particular case involve a complex anal-ysis of the patients condition and available courses ofaction.Therefore,clinical considerations may lead anendoscopist to take a course of action that varies from thisposition statement.Since its introduction in 1968,ERCP has become a com-monly performed endoscopic procedure.2The diagnosticand therapeutic utility of ERCP has been well demonstratedfor a variety of disorders,including the management of cho-ledocholithiasis,the diagnosis and management of biliaryand pancreatic neoplasms,and the postoperative manage-ment of biliary perioperative complications.3-5The evolutionof the role of ERCP has occurred simultaneously with that ofother diagnostic and therapeutic modalities,most notablymagnetic resonance imaging/MRCP,laparoscopic cholecys-tectomy(with or without intraoperative cholangiography),and EUS.For endoscopists to accurately assess the clinicalappropriateness of ERCP,it is important to have a thoroughunderstanding of the potential complications of this proce-dure.Numerousstudieshavehelpeddeterminetheexpectedrates of complications,potential contributing factors for theseadverse events,and possible methods for improving thesafety of ERCP.Recognition and understanding of potentialcomplications of ERCP are vital in the acquisition of appro-priate informed consent.6Reported complication rates varywidely in the published literature because of differences instudy design,patient population,and definitions of compli-cations.The diagnosis and management of all complicationsof ERCP are beyond the scope of this document;however,general principles are discussed.PANCREATITISIncidencePancreatitisisthemostcommonseriousERCPcomplication.7-15Although transient increase in serumpancreatic enzymes may occur in as many as 75%ofpatients,16such an increase does not necessarily constitutepancreatitis.A widely used consensus definition for post-ERCP pancreatitis(PEP)is(1)new or worsened abdominalpain,(2)new or prolongation of hospitalization for at least2 days,and(3)serum amylase 3 times or more the upperlimit of normal,measured more than 24 hours after theprocedure.17By using this or similar definitions,the inci-dence of PEP in a meta-analysis of 21 prospective studieswas approximately 3.5%18but ranges widely(1.6%-15.7%)depending on patient selection.19,20The rates of PEP inpediatric patients approach those seen in adults.21Risk factorsNumerous factors have been found to correlate with thedevelopment of PEP.Some of these are patient specific(eg,age,sex,history of PEP),whereas others are related tothe procedure itself(eg,pancreatic sphincterotomy,precutsphincterotomy)or endoscopist experience.Risk factorsfor PEP that have been studied in large,prospective mul-tivariate analyses are summarized in Table 1.22Risk factorscan be synergistic.For example,Freeman et al9demon-strated that the risk of pancreatitis in a female with anormal bilirubin level and suspected sphincter of Oddidysfunction(SOD)is 18%compared with 1.1%for a typ-ical low-risk patient.Risk of PEP associated with the use ofa precut or access papillotomy is controversial.Factorssuch as endoscopist experience and timing of precut mayaffect the risk,although the literature is mixed.7,8,10,23-29ERCP in the setting of suspected SOD is associated withCopyright 2012 by the American Society for Gastrointestinal Endoscopy0016-5107/$36.00doi:10.1016/j.gie.2011.07.010www.giejournal.orgVolume 75,No.3:2012GASTROINTESTINAL ENDOSCOPY467increased risk of pancreatitis(as high as 20%-25%),irre-spective of whether manometry is performed.9When per-formed with aspiration-type catheters,manometry was notassociated with an incremental increased risk of pancre-atitis in multivariate analysis.9,30Endoscopic papillaryballoon dilation has been proposed as an alternative toendoscopic biliary sphincterotomy;however,2 meta-analyses have shown a statistically significant increasedrisk of PEP with endoscopic papillary balloon dilationcompared with standard sphincterotomy.31,32Methods of reducing post-ERCP pancreatitisRecognition and understanding of risk factors for PEPhave allowed endoscopists to provide a more accurateestimate of an individuals risk of PEP and to direct pre-ventive measures in appropriate clinical situations.Patient selection.Appropriate patient selection is in-strumental in reducing PEP.Other imaging modalitiesshould first be considered for the diagnosis of commonbile duct stones and pancreaticobiliary malignancy.Manyof the variables identified in multivariate analyses(Table1)can be assessed pre-procedure and should be ac-counted for when considering ERCP.In general,alterna-tives to ERCP should be considered when multiple riskfactors are present and the likelihood of therapeutic inter-vention is low.MRCP and EUS both have sensitivity similar to that ofERCP for the detection of many pancreaticobiliary disor-ders without the associated risk of pancreatitis.3,33-35ERCPshould be reserved for those patients with a reasonablelikelihood of requiring therapeutic intervention,based oneither clinical criteria(eg,cholangitis)or abnormalitiesidentified by other imaging modalities.Pharmacologic prophylaxis.Several agents for thepharmacologic prophylaxis of PEP have been proposed,each directed toward the interruption or amelioration ofsome aspect of the inflammatory cascade that accompa-nies and potentiates acute pancreatitis.Meta-analyses haveshown a statistically significant reduction of PEP with in-domethacin or diclofenac given rectally just before ERCPor on arrival at the recovery room.36-38Many studies werelimited to high-risk patients.However,other studies oforal nonsteroidal anti-inflammatory drugs have shown nobenefit.39Nitroglycerin was shown to reduce the incidenceof PEP in 2 meta-analyses,but methodologic limitationsand the side-effect profile of nitroglycerin hinder it frombeing recommended in the prevention of PEP.40-42Othermeta-analyses have found no benefit from somatostatin,octreotide,or low osmolality contrast for the prevention ofPEP.42,43Finally,additional studies have shown that glu-cocorticoids and gabexate are ineffective in the preventionof PEP.44-46Modifications in technique to prevent pancreati-tis.Pancreatic duct stents.Multiple prospective studieshave shown the benefits of temporary pancreatic duct(PD)stents in lowering the risk and severity of PEP inhigh-risk populations,such as those undergoing SOD ma-nometry,ampullectomy,pancreatic sphincterotomy,pre-cut sphincterotomy,pancreatic brush cytology,difficultbiliary cannulation,and manipulation of the PD withwires.47,48In a systematic review involving 680 patients in8 studies,pancreatitis was significantly reduced with PDstents from 19%in controls to 6%.The number needed totreat to avoid a single episode of PEP with PD stentplacement was 8.49A cost-effectiveness analysis suggestedthat PD stent placement in high-risk patients may be cost-effective for the prevention of PEP.50Wire-guided cannulation.The use of wire-guidedcannulation before contrast injection has been shown inmeta-analyses to result in greater success of biliary cannu-lation and lower risk of PEP by avoiding the injection ofcontrast into the pancreas.51,52Data are mixed as towhether inadvertent wire-guided cannulation of the PD isan independent risk factor for PEP.53,54Electrocautery setting.A meta-analysis of 4 studiescomparing pure-cut current versus blended current in pa-tients undergoing endoscopic biliary sphincterotomy dem-onstrated no statistically significant difference in the rate ofPEP.55HEMORRHAGEMost ERCP-associated bleeding is intraluminal,althoughintraductal bleeding can occur and hematomas(hepatic,splenic,and intra-abdominal)have been reported.56-58Hem-orrhage is primarily a complication related to sphincterotomyrather than diagnostic ERCP.In a meta-analysis of 21 pro-spective trials,the rate of hemorrhage as a complication ofERCP was 1.3%(95%CI,1.2%-1.5%)with 70%of thebleeding episodes classified as mild.18Hemorrhagic com-plications may be immediate or delayed,with recognitionoccurring up to 2 weeks after the procedure.The risk ofsevere hemorrhage(ie,requiring?5 units of blood,sur-TABLE 1.Risk factors for post-ERCP pancreatitis inmultivariate analysesBalloon dilation of biliary sphincterHistory of post-ERCP pancreatitisNormal bilirubinPancreatic duct injectionPancreatic sphincterotomyPrecut sphincterotomySuspected sphincter of Oddi dysfunctionYoung ageModified from Freeman.22Complications of ERCP468GASTROINTESTINAL ENDOSCOPYVolume 75,No.3:2012www.giejournal.orggeryor angiography)is estimated to occur in fewer than1 per 1000 sphincterotomies.59Although sphincterotomy alone is a risk factor for hem-orrhage,other factors identified in multivariate analysisinclude coagulopathy,the use of anticoagulants within 72hours of sphincterotomy,the presence of acute cholangitisor papillary stenosis,the use of precut sphincterotomy,and low case volume of the endoscopist(ie,1 sphincter-otomy per week or less).7,8,10Observed bleeding duringthe initial examination is also predictive of delayed bleed-ing.7Neither the length of incision nor the preprocedureuse of aspirin or other nonsteroidal anti-inflammatorydrugs appear to be important predictors of bleeding.7Alarge,multicenter study of 4561 patients undergoing ERCPfound that the risk of post-ERCP hemorrhage was associ-ated with hemodialysis,visible bleeding during the proce-dure,higher bilirubin,and the use of pure-cut current forsphincterotomy.14Antiplatelet treatment,precut sphincter-otomy,coagulopathy,and cholangitis were not associatedwith post-ERCP hemorrhage.The use of a microprocessor-controlled ERBE electrosurgical generator for sphincterot-omy has been associated with a lower rate of endoscopi-cally visible bleeding,but not clinically evident bleedingcompared with conventional electrocautery.60More de-tailed data on the safety of various types of current areneeded.Treatment of bleeding includes injection therapywith epinephrine,with or without thermal therapy,andendoscopic clips.61ERCP with sphincterotomy is consid-ered a higher risk procedure for bleeding,and antithrom-botic therapy should be adjusted according to publishedguidelines.62PERFORATIONPerforation rates with ERCP range from 0.1%to0.6%.7,8,10,15,63Three distinct types of perforation havebeen described:guidewire-induced perforation,periamp-ullary perforation during sphincterotomy,and luminal per-foration at a site remote from the papilla.63Risk factors forperforation determined in a large retrospective study in-cluded the performance of a sphincterotomy,Billroth IIanatomy,the intramural injection of contrast,prolongedduration of procedure,biliary stricture dilation,andSOD.10,64However,in a more recent multicenter prospec-tive study,only malignancy and precut access were asso-ciated with an increased risk of perforation.14Prompt rec-ognition of periampullary perforation and treatment withaggressive biliary and duodenal drainage(by means ofnasobiliary and nasogastric tubes)coupled with broad-spectrum antibiotics can result in clinical resolution with-out the need for operative intervention in as many as 86%of patients.64The management of perforation will depend on manyfactors,such as the site and location,clinical status,andradiographic imaging.Early identification and expeditiousmanagement of a perforation have been shown to de-crease associated morbidity and mortality.65Perforationsrelated to endoscopy are best approached in collaborationwith surgical colleagues.INFECTIONCholangitisThe rate of post-ERCP cholangitis is 1%or less.7,8,10Riskfactors identified as significant include the use of com-bined percutaneous-endoscopic procedures,stent place-ment in malignant strictures,the presence of jaundice,primary sclerosing cholangitis,low case volume,and in-complete or failed biliary drainage.7In the case of malig-nant hilar obstruction(ie,Klatskin tumor),it is suggestedthat endoscopists avoid filling all intrahepatic segmentsand drain all intrahepatic segments that are filled withcontrast.66Unilateral endoscopic biliary stent placementdirected by previous imaging(eg,MRCP)has been shownto offer palliation of jaundice equal to bilateral placementbut with less risk of cholangitis.66-68In a study of 188patients with inoperable malignant hilar obstruction,post-ERCP cholangitis rates were lower in patients undergoingair cholangiography(3%)compared with those who hadtraditional iodine contrast studies before stenting(24%).69CholecystitisCholecystitis complicates approximately 0.2%to 0.5%of ERCPs.7,8The risk appears to be correlated with thepresence of stones in the gallbladder and possibly filling ofthe gallbladder with contrast during the examination.7Additionally,placement of self-expandable metal stentsmay increase the risk of cholecystitis,particularly if thestent is covered and the cystic duct is obstructed.70-72Prevention of infectionProphylactic antibiotics.Two meta-analyses failed toshow the benefit of routine prophylactic antibiotic use inERCP.73,74A recent retrospective analysis of 11,484 ERCPsover an 11-year time period at a single institution assessedthe role of antibiotics in preventing cholangitis.Althoughthe use of routine prophylactic antibiotics decreased from95%of patients to 25%,the infection rate decreased from0.48%to 0.25%.Multivariate analysis indicated that onlytransplant recipients undergoing biliary intervention werefound to be at increased risk of infection.76ASGE guidelines currently recommend that antibioticprophylaxis should be considered before an ERCP in pa-tients with known or suspected biliary obstruction inwhich there is a possibility that