Management of patients on antithrombotic therapy prior to gastrointestinal endoscopy
The editors would like to acknowledge Bernard Boneu, Luc Maillard, Charles-Marc Samama, Jean-François Sched, Gérard Gay, Thierry Ponchon, Denis Sautereau, Jean-Pierre Arpurt, Christian Boustière, Jean Boyer, Jean Cassigneul, Pierre-Adrien Dalbies, Jean Escourrou, René Laugier, Runo Richard-Molard, Gilbert Tucat, Bruno Vedrenne.
Procedure-related bleeding 53
Bleeding associated with antithrombotic therapy 53
Risks associated with discontinuation of antithrombotic therapy 54
Switching from warfarin or antiplatelet drugs to alternative therapy 56
Resumption of antithrombotic treatment 56
Guidelines for the management of antithrombotic therapy are important but decisions must be made on a case-by-case basis.
Bleeding is common but rarely life-threatening, while thromboembolism resulting from interruption of antithrombotic therapy may be fatal or permanently disabling.
Low risk procedures can be undertaken in anticoagulated patients as long as treatment is within the therapeutic range.
High-risk procedures that are non-urgent should be deferred until antithrombotic therapy has been completed or until such time as the risks of stopping therapy are low.
For most procedures, aspirin therapy does not need to be stopped.
If in doubt, the cardiologist or physician who prescribed antithrombotic therapy should be consulted for advice before any planned procedure.
Where possible, low-risk procedures (e.g. biliary stenting without sphincterotomy) and measures to minimize bleeding (clips, loops), should be considered.
The management of patients taking anticoagulant or antiplatelet drugs is an increasingly common problem prior to endoscopic procedures. The risk of bleeding must be given serious consideration, but without underestimating the risk of discontinuing antithrombotic treatment: bleeding is rarely life-threatening, whereas a thromboembolic event associated with interruption of therapy may be permanently disabling or fatal. It is not possible to formulate guidelines that cover all conceivable clinical scenarios and management of a patient’s antithrombotic therapy must be individualized.
Ideally, any necessary therapeutic adjustments should be undertaken in consultation with the doctor who prescribed the treatment in question, and based on a case-by-case evaluation of the risk–benefit ratio of the planned procedure and the envisaged changes in the patient’s antithrombotic therapy. Unfortunately, in most cases, these ideal conditions do not pertain, for a number of reasons: the patient may have forgotten which doctor prescribed the drug or the doctor may be unavailable; or there may be a difference of opinion or lack of knowledge regarding the respective risks entailed by the procedure and discontinuation of the medication. In deciding what to do, the following factors should be taken into account:
The bleeding risk associated with the procedure
The antithrombotic therapy being used
The thromboembolic risk associated with discontinuation of therapy.
In many countries endoscopists, cardiologists, and hematologists have developed consensus recommendations, which of course are only indicative and may evolve over time, as new information emerges. Several of these guidelines are listed at the end of this chapter.
There is little information available in the literature concerning the hemorrhagic risk associated with antithrombotic agents. It is reasonable to presume that such therapy increases the risk of bleeding above normal levels by rendering bleeding episodes symptomatic that would go unnoticed in the presence of normal coagulation. However, antithrombotics do not themselves cause bleeding.
Hence, antithrombotics have little impact on procedures whose bleeding risk is low. The following factors should be taken into consideration in assessing the risk of a procedure that is performed on a patient who is taking an antithrombotic:
In the case of procedures for which the risk of spontaneous bleeding is >1%, antithrombotic therapy may exacerbate bleeding. Hence, the risk of bleeding associated with a sphincterotomy would increase by a factor of 7.8 on warfarin therapy.
The incidence (but not severity) of bleeding may increase for bleeding that is inaccessible to endoscopic hemostasis, e.g. endoscopic ultrasound-guided FNA.
In the presence of an elevated risk of perforation, antithrombotic therapy may complicate management should this occur.
Procedure risk may be classified as below.
Low-risk procedures ( Box 1 )
Colonoscopy without polypectomy.
Diagnostic endoscopic ultrasound (no FNA /biopsy).
ERCP (with ampullary or biliary dilatation, or insertion of a stent without sphincterotomy).
Enteral stent placement (without dilatation).
In these, bleeding occurs only rarely and can be managed endoscopically. In the case of colonoscopy, it is rarely known in advance whether there will be polyps requiring resection. However, the perforation risk entailed by the examination may increase the risks of maintaining antithrombotic therapy if emergency surgery turns out to be necessary. These two factors may sometimes prompt classification of colonoscopy as a risky procedure, as a precautionary measure. Taking a mucosal biopsy at these procedures (regardless of whether standard or pediatric forceps are used) does not significantly increase the risk of bleeding.
High risk of bleeding
Table 1 summarizes the estimated risks of bleeding with various endoscopic procedures (1% or more), in settings where the bleeding can be managed endoscopically.
|Procedure||Estimated bleeding risk (%)|
|Gastric polypectomy or jumbo/snare biopsy||4|
|Endoscopic mucosal resection||≤22|
|Endoscopic treatment of esophageal or gastric varices||≤6|
|Endoscopic hemostasis of vascular lesions||≤5|
Low risk of bleeding (<1%) in settings where bleeding cannot be monitored endoscopically
Endoscopic ultrasound-guided FNA
Gastrointestinal stricture dilatation by bougie or balloon dilator. Insertion of a self-expanding metallic stent without dilatation
Transnasal gastroscopy (risk of epistaxis).
Bleeding associated with antithrombotic therapy
These drugs inhibit platelet function, particularly activation and aggregation.
Aspirin and most NSAIDs inhibit platelet aggregation. Aspirin induces irreversible inhibition of cyclooxygenase. Doses usually range from 75 to 325 mg/day (1–2 mg/kg in practice). Disturbance in clotting is not fully corrected until all platelets have been replaced, which takes 7–10 days.
However, a functional platelet level of 50 × 10 9 /L is regarded as adequate for normal hemostatic function. Platelets are renewed at a rate of 10% per day. Thus, depending on the baseline platelet level, discontinuing treatment for 3–5 days is generally sufficient for a patient to recover their normal hemostatic function. NSAIDs also inhibit cyclooxygenase, albeit reversibly. The duration of action is temporary and is determined by the individual drug’s half-life.
The limited data available suggest that standard doses of aspirin and NSAIDs do not significantly increase the risk of bleeding secondary to endoscopic biopsy, colonic snare polypectomy or biliary sphincterotomy. For snare polypectomy, use of a detachable loop is recommended in the presence of a polyp stalk >1 cm. There are no data concerning polyp resection by endoscopic mucosal resection (EMR) or other high-risk procedures in patients on aspirin therapy.
Thienopyridines, dipyridamole, and glycoprotein inhibitors
Thienopyridines (ticlopidine and clopidogrel) irreversibly suppress one of the adenosine diphosphate (ADP) platelet receptors. These preparations frequently prolong the bleeding time, but there is currently no test available for measurement of restoration of normal platelet function. In cases where therapy is suspended, it is recommended that 7–10 days be allowed for platelet renewal. However, as is the case with aspirin, 3–5 days may be sufficient for the return of normal hemostatic function (see above). Platelet transfusion can be beneficial for bleeding management. Ticlopidine has been largely replaced by clopidogrel because of side-effects. There is also little data available on the newer thienopyridine prasugrel. Dipyridamole has only a moderate antiplatelet effect and does not increase the risk of bleeding. Glycoprotein IIb/IIIa inhibitors, abciximab, eptifibatide, and tirofiban, are used parenterally in the short-term management of acute coronary syndromes. Endoscopy in patients receiving these agents would only occur in life-threatening situations and after careful discussion with the patient’s cardiologist, and these drugs are not the subject of this chapter.
Vitamin K antagonists
Coumarin (warfarin, acenocoumarol) and phenindione are the only oral anticoagulant drugs available today. Their plasma half-lives vary but normal coagulation is usually restored 2–4 days following discontinuation of the drug. The anticoagulation effect achieved depends on the dose used and the patient’s susceptibility to the drug. The International Normalized Ratio (INR) accurately reflects this effect. The risk of bleeding increases as the INR rises, an INR ranging from 2 to 3 is generally needed for effective anticoagulation and carries a low risk of bleeding.
An INR of 3–4.5 is necessary in the presence of a major thromboembolic risk. If the INR is <1.5, the bleeding risk is on a par with that of non-treated patients. In the event of an overdose (INR >6) and/or bleeding, various options are available. In the absence of bleeding, discontinuation of treatment and oral vitamin K intake may be sufficient. In the presence of major bleeding, intravenous vitamin K (5–10 mg) along with IV prothrombin complex, 30–50 units/kg or fresh frozen plasma should be considered. For minor bleeding, smaller doses of IV vitamin K should be given, e.g. 0.5–2.5 mg.
Giving large doses of vitamin K can make it difficult to re-establish anticoagulation with warfarin after the bleeding episode is resolved. Discuss with Hematology and the patient’s cardiologist/specialist before giving IV vitamin K.
Heparins are the third family of antithrombotics that are widely used. The main indications for heparins are prophylaxis and therapy of venous thromboembolic disease, including deep venous thrombosis (DVT), pulmonary embolism, acute coronary syndromes, and thrombosis prophylaxis in patients with mechanical cardiac valve prostheses before switching to oral anticoagulation. Only low-molecular weight heparins (LMWHs) are used for DVT prophylaxis, in doses ranging from 2000 to 5000 U once daily according to the thrombosis risk level and the product used. The bleeding risk with these doses is low, becoming negligible 12 hour following administration. For other indications, either subcutaneous LMWHs or unfractionated heparin administered intravenously is used.
The LMWH doses in such settings vary considerably depending on the preparation used and body weight. If a LMWH is administered twice daily, it is necessary to wait 12–18 hour for the heparin level to return to a level that allows normal hemostasis. If the drug is administered once daily, 24 hour must be allowed. It should be noted that the activated partial thromboplastin time (APTT) is not useful in such cases as it remains normal, even in the presence of persistent heparinemia. The only way to verify the presence of residual heparinemia is the anti-Xa activity level, which should be <0.20 U/mL. The half-life for unfractionated heparin administered intravenously at a dose of 400–600 U/kg per day is 45–90 minutes and normal hemostasis returns 4–6 hour after the infusion ends. In these cases, the APTT may be useful for verifying that clotting has normalized. If unfractionated heparin is administered via two or three injections using the same dose, because of the longer half-life, it is necessary to wait 8–12 hour for normal coagulation to return. Here too the APTT may allow verification of this.
New antithrombotics are becoming available: an anti-Xa preparation known as fondaparinux (Arixtra) is available and is used for prophylaxis of deep venous thrombosis following orthopedic surgery. The bleeding risk is about the same as with LMWHs but there is no known antidote in the event of bleeding. Fondaparinux has a half-life of 15 hour and endoscopy should not be performed until the drug has been completely eliminated, which takes 4–5 half-lives.
Risks associated with discontinuation of antithrombotic therapy
The risk associated with discontinuation of antithrombotic therapy ranges from minor to major, depending on the specific indications involved. Cases of sudden death or coronary stent occlusion within 7 days of discontinuation have been described.
Patients receiving oral anticoagulants
Indications associated with acute thromboembolic risk ( Table 2 )
When taking a patient off warfarin, it is essential to use a specific protocol ( Box 2 ) involving unfractionated heparin. It is also advisable to carefully weigh whether the planned procedure is truly indicated, and if so, undertake procedures associated with a low risk of bleeding (e.g. insertion of a biliary stent without sphincterotomy).
Switching medications in patients with acute thromboembolic risk
Discontinuation of warfarin 3–5 days prior to the procedure.
The day after discontinuation, begin an anticoagulant dose of heparin; monitor the patient’s APTT.
Check INR the day before the examination.
Discontinue unfractionated IV heparin 4–6 hour prior to endoscopy or the final injection of LMWH 8 hour (three injections/day) or 12 hour before the procedure (two injections/day); resume heparin 6–8 hour post-procedure or immediately if there is no risk of bleeding.
Resume warfarin the same evening (or later, see above).
Discontinue heparin once two daily INR results are satisfactory.
Switch to unfractionated heparin (or, in some countries, LMWH): same as for acute thromboembolic risk.
Discontinue warfarin 3–5 days before the procedure.
Begin therapeutic doses of LMWH the following day.
Discontinue LMWH on the day before the procedure and check the INR and platelets.
Resume LMWH 12 hour post-procedure.
Resume warfarin the same evening (or later, see above).
Discontinue LMWH if two INR results 2 days apart are between 2 and 3.
Taking patients off warfarin is usually undertaken using therapeutic doses of IV unfractionated heparin (400–600 IU/kg per 24 hour), which remains the reference method.
|All prosthetic metal valves in a mitral position||3–4.5|
|All first-generation metal aortic valves||3–4.5|
|Second-generation aortic valves in patients with an additional embolic risk factor||3–4.5|
|Atrial fibrillation associated with other thromboembolic risk factors, particularly mitral valve disease||2–3|
Indications associated with low or moderate thromboembolic risk ( Box 3 )
Coronary syndromes and/or DVT management.
Atrial fibrillation in patients over 65 years or with associated thromboembolic risk factors (history of CVA, cardiac failure <2 months previously or left ventricular dilatation with end diastolic dimension >60 mm).
Some cases of mitral regurgitation or stenosis (with a dilated left-atrium).
Mechanical aortic valves in the absence of embolic risk factors.
Temporary discontinuation of antithrombotic therapy is usually sufficient in these situations.
Patients receiving antiplatelet drugs
Indications associated with acute thromboembolic risk ( Box 4 )
Switching medications in patients with acute thromboembolic risk
In view of the high risk associated with the discontinuation of antiplatelet drugs and the brevity of the treatment period, high-risk procedures should be postponed if possible. If the procedure is essential, the patient’s medication should only be switched after consulting with their specialist and either non-fractionated heparin or LMWH in therapeutic doses is used.
Switching to a LMWH:
Discontinue antiplatelet agent according to the drug’s half-life (7–10 days) or the time needed to restore adequate function (i.e. 50 × 10 9 /L functional platelets, where platelets are renewed at a rate of 10% per day of the baseline platelet count).
Begin therapeutic dose LMWH on the following day.
Discontinue LMWH 24 hour prior to procedure and check coagulation screen.
Resume antiplatelet agents the day post-procedure (or later; see above).
Acute coronary syndromes that began <4 weeks previously
Insertion of a bare coronary stent <1 month previously
Insertion of a drug-eluting coronary stent <2 months previously for sirolimus and <6 months previously for paclitaxel. If placed >12 months ago, the risks of stent occlusion from stopping antiplatelet therapy are much lower
Endocoronary radiotherapy <1 year previously.
Bare (uncoated) metal coronary stents require 1 month for re-endothelialization, while this process may take at least 6 months for drug-eluting stents. Until this occurs, there is a 50% risk of acute myocardial infarction or death. A patient should only be taken off antiplatelet drugs after consulting the cardiologist or other specialist responsible for the patient and where possible, therapy should be restarted within 5 days as the thrombosis risk increases after this time. Unfractionated heparin alone or therapeutic dose LMWH should be used. Again, low-risk procedures should be used where possible if endoscopy cannot be deferred.
Indications associated with moderate thromboembolic risk
Acute coronary syndromes that began >4 weeks previously; stable angina
Secondary prophylaxis for myocardial or arterial infarction
Secondary prophylaxis for CVA in the absence of a source of cardiac embolus
Atrial fibrillation in patients <65 years old with no thromboembolic risk factors or source of cardiac embolus
Peripheral vascular disease.
Temporary discontinuation of antithrombotic therapy is usually sufficient.
Indications associated with minor thromboembolic risk
The main goal here is primary prevention of death or myocardial infarction in patients over the age of 50 with at least one vascular risk factor. Discontinuation of the therapy may be considered.
Switching from warfarin or antiplatelet drugs to alternative therapy
No drug has been approved for switching from warfarin or antiplatelet therapy. The discontinuation/switching procedure takes account of the treatment currently being used and the patient’s thromboembolic risk factors ( Boxes 2 and 4 ).
Calcium heparin may also be effective. The total required dose per 24 hour (in units) is the same as or slightly higher than the dose needed for continuous infusion. The drug can be administered by two or three daily subcutaneous injections. The efficacy of therapy is assessed by the APTT. The target APTT ratio should be 2–3. If heparin levels are measured to guide results, they should be between 0.3 and 0.6 IU/L. If there is a discrepancy between the unfractionated heparin dose injected and the resulting APTT results, measurement of circulating heparin levels is recommended.
The efficacy of switching to LMWH has been validated but guidelines may vary from country to country and local practice needs to be clarified and followed. If LMWHs are used, the dose must be adjusted to the patient’s weight and should be given by one or two subcutaneous daily injections.
Heparin-induced thrombocytopenia with or without thrombosis (HITT) is a rare but serious complication of treatment with either unfractionated heparin or LMWHs and should be suspected if the platelet count drops by 50% or more. If this occurs, all heparins should be stopped and an alternative thrombin inhibitor (e.g. danaparoid) used.
A patient can be taken off antiplatelet drugs using preparations with short-term and reversible anti-thrombotic action. However, no treatment of this kind has been validated prospectively.
LMWHs are an alternative and are administered using the same protocol as for switching to warfarin.
Resumption of antithrombotic treatment
While most bleeding occurs within 24 hour postoperatively, it can also be delayed (for up to 3 weeks following polypectomy). Starting the patient on antithrombotic therapy too soon can provoke delayed bleeding. Following a sphincterotomy, the risk of bleeding is put at 10–15% if warfarin is resumed before the third post-procedure day. There is little hard data to guide when to resume antithrombotic therapy and each case must be decided on its individual merits.
In cases where antithrombotic therapy is temporary and an endoscopic procedure would carry a high risk of bleeding and is not urgent, the procedure should be postponed.
In an emergency setting or if antithrombotic therapy was not discontinued and/or is scheduled to resume at an early stage, the following, which reduce the risk of bleeding, should be used wherever possible: detachable loops; hemostatic clips; pancreatic or biliary stents without a sphincterotomy.
All equipment needed for endoscopic hemostasis should be ready to hand. Use of drugs or blood products that inhibit antithrombotic action should also be considered in light of the endoscopic alternatives and the relative thromboembolic risk.
If an emergency situation does not allow for determination of a patient’s thromboembolic risk, as a precaution the patient should be regarded as high risk.
Setting-specific therapy adjustments
Elevated risk of bleeding (1% or more), if the patient can be treated endoscopically ( Table 3 ):
In the absence of any other hemostatic abnormality, colonoscopy with snare polypectomy and sphincterotomy can be performed without stopping standard-dose aspirin or NSAIDs. A detachable loop should be used for colonic polypectomies if the stalk diameter is >1 cm. Mucosectomy (EMR) should not be performed while the patient is taking aspirin.
Apart from these special situations, antithrombotics should be discontinued pre-procedure, for as long as necessary to deactivate them. For warfarin, confirmation that the INR has normalized should be sought.
|Antiplatelet drugs||Vitamin K antagonists||New antithrombotics|
|Flexible sigmoidoscopy with or without biopsy||Yes||Recent INR||Yes|
|Colonoscopy without polypectomy, with or without biopsy||Yes||Recent INR||Yes|
|ERCP without sphincterotomy, with or without biopsy||Yes||Recent INR||Yes|
|Enteroscopy with or without biopsy||Yes||Recent INR||Yes|
|ERCP with endoscopic sphincterotomy||Yes||No||No|
|Gastric strip biopsy and polypectomy||No||No||No|
|Photodynamic therapy and radiofrequency ablation||No||No||No|
|Management of esophageal or gastric varices||No||No||No|
|Hemostasis for vascular lesions||No||No||No|
|Endoscopic ultrasound-guided FNA||No||No||No|
|Metallic stent insertion without dilatation||No||No||No|
Low risk of bleeding (<1%) in patients who cannot be monitored endoscopically:
All antithrombotics should be discontinued prior to the procedure, for as long as necessary to deactivate them.
For warfarin, again the INR needs to be checked pre-procedure.
No adjustment of anticoagulant or antiplatelet therapy is necessary. For warfarin, a recent verification of the INR is advisable. Non-urgent procedures should be postponed if anticoagulation is outside the therapeutic range.
Treatment discontinuation and resumption
Anticoagulant treatment should be discontinued in accordance with the patient’s thromboembolic risk, as follows:
In patients at high risk During the discontinuation period, the patient is put on a suitable antithrombotic and monitored.
In patients at moderate risk During the discontinuation period, the possibility of putting the patient on a suitable antithrombotic is considered; the patient is monitored if an antithrombotic is administered.
In patients at low risk No discontinuation of therapy is necessary.
Antithrombotic therapy is resumed post-procedure. In view of the risk of delayed bleeding, the benefits of immediate resumption of antithrombotic therapy should be weighed against the increased risk of bleeding and each case must be considered on its own merits.
It is difficult to formulate recommendations that will cover all possible clinical scenarios that endoscopists will encounter. The bleeding risk associated with the procedure and the underlying pathology (thromboembolic risk) must be assessed on a case-by-case basis. Clear communication between the endoscopist and the prescribing physician is important in deciding the best strategy. The guidance provided in this chapter can potentially serve as a basis for such discussions but it should be borne in mind that this consensus is somewhat arbitrary in places because of lack of an evidence base for recommendations and also because guidelines differ from country to country. The British Society of Gastroenterology guidelines, for example, include a simple, practical flowchart that is a useful starting point for many of the common clinical situations ( Fig. 1 ).
Antibiotics for the prevention of infective endocarditis 59
Antibiotic prophylaxis in specific groups of patients 60
Antibiotic prophylaxis for specific endoscopic procedures 60
Prophylactic antibiotics are no longer recommended purely as prophylaxis against infective endocarditis.
Patients with cirrhosis who present with an upper GI bleed should be given prophylactic antibiotics.
Patients who have biliary obstruction with sepsis, or who have biliary obstruction which remains undrained following ERCP, should remain on antibiotics until the biliary obstruction is relieved.
Patients who undergo a EUS-FNA of a cystic lesion should be given prophylactic antibiotics.
All patients who have a PEG inserted should have pre-procedure antibiotics.
Patients undergoing EUS or ERCP drainage of a pseudocyst require prophylactic antibiotics.
Previously, antibiotics have been prescribed for three indications associated with endoscopic procedures:
To prevent infective endocarditis in patients with heart disease
To treat patients with infection prior to endoscopy (e.g. ascending cholangitis)
As prophylaxis against infection following an endoscopic procedure.
European and American Society guidelines ( Table 1 ) have recently changed significantly with respect to antibiotic prophylaxis against infective endocarditis. Antibiotics are no longer recommended for gastrointestinal procedures in the absence of established infection, but are still recommended for patients with evidence of infection prior to endoscopy and in patients undergoing specific procedures, which are discussed below.
It is important to have knowledge of local antibiotic resistance.
Society guidelines for antibiotic prophylaxis vary. Ensure that you are aware of your own local guidelines.
|AHA 2007||Prophylaxis against infective endocarditis is not recommended for non-dental procedures such as transesophageal echocardiogram, EGD or colonoscopy in the absence of active infection.||;|
|NICE 2008||Antibiotic prophylaxis for gastrointestinal procedures is not recommended.|
|ESC||Antibiotic prophylaxis is not recommended for gastroscopy, colonoscopy, or transesophageal echocardiography.|
|BSG||Antibiotics are not indicated as prophylaxis against infective endocarditis.|
|ASGE||Antibiotic prophylaxis for infectious endocarditis is not recommended.|
Antibiotics for the prevention of infective endocarditis
European and American societies have recently significantly altered their recommendations for endocarditis prophylaxis in patients undergoing endoscopy (see Further Reading ). The rationale for these changes has been summarized by the European Society of Cardiology:
The incidence of bacteremia after endoscopic procedures versus the risk of bacteremia associated with routine daily activities.
The incidence of transient bacteremia varies greatly in studies. Transient bacteremia occurs frequently in daily routine activities such as brushing teeth, flossing or chewing. It therefore appears that a large proportion of cases of infective endocarditis arise from these daily activities and are not procedure related.
Risks and benefits of prophylaxis:
The procedure-related risk of a patient developing infective endocarditis ranges from 1 : 14 000 000 for dental procedures in the average population to 1 : 95 000 for patients with previous infectious endocarditis. These estimates demonstrate the huge number of patients that require treatment to prevent a single case of infective endocarditis.
In the majority of patients, no procedure preceded the development of infective endocarditis. Therefore, infective endocarditis prophylaxis may at best protect a small proportion of patients, while the bacteremia that causes infective endocarditis in the majority of patients appears to derive from another source.
Antibiotic administration carries a small risk of anaphylaxis.
Widespread or inappropriate use of antibiotics can lead to resistant microorganisms.
Lack of scientific evidence for the efficacy of infective endocarditis prophylaxis:
Studies reporting on the efficacy of antibiotic prophylaxis to prevent or alter bacteremia are contradictory, with no data demonstrating that reduced duration or frequency of bacteremia after any medical procedure leads to a reduced procedure-related risk of infective endocarditis.
The efficacy of antibiotic prophylaxis itself has never been investigated in a prospective randomized controlled trial and assumptions of efficacy are based on non-uniform expert opinion, animal experiments, case reports and contradictory observational studies.
All societies recommend that in patients with an established infection prior to endoscopy, or in patients undergoing a procedure which will require antibiotics (see below), antibiotic coverage for enterococci should be given (i.e. amoxicillin, ampicillin, piperacillin or vancomycin) in addition to the antibiotics for the specific indication.
Antibiotic prophylaxis in specific groups of patients
The guidelines of the American Society of Gastrointestinal Endoscopy can be found in Table 2 .
|Patient condition||Procedure contemplated||Goal of prophylaxis||Peri-procedural antibiotic prophylaxis||Grade of recommendation; comments|
|All cardiac conditions||Any endoscopic procedure||Prevention of infective endocarditis||Not indicated||1C+|
|Bile-duct obstruction in the absence of cholangitis||ERCP with complete drainage||Prevention of cholangitis||Not recommended||1C|
|Bile-duct obstruction in the absence of cholangitis||ERCP with anticipated incomplete drainage (e.g. PSC, hilar strictures)||Prevention of cholangitis||Recommended; continue antibiotics after the procedure||2C|
|Sterile pancreatic fluid collection (e.g. pseudocyst, necrosis), which communicates with pancreatic duct||ERCP||Prevention of cyst infection||Recommended||3|
|Sterile pancreatic fluid collection||Transmural drainage||Prevention of cyst infection||Recommended||3|
|Solid lesion along upper-GI tract||EUS-FNA||Prevention of local infection||Not recommended||1C; low rates of bacteremia and local infection|
|Solid lesion along lower-GI tract||EUS-FNA||Prevention of local infection||Insufficient data to make firm recommendation||Endoscopists may choose on a case-by-case basis; a single study indicates a low risk of infection|
|Cystic lesions along GI tract (including mediastinum)||EUS-FNA||Prevention of cyst infection||Recommended||1C|
|All patients||Percutaneous endoscopic feeding tube placement||Prevention of peristomal infection||Recommended||1A; decreases risk of soft tissue infection|
|Cirrhosis with acute GI bleeding||Required for all patients, regardless of endoscopic procedures||Prevention of infectious complications and reduction of mortality||Upon admission||1B; risk for bacterial infection associated with cirrhosis and GI bleeding is well established|
|Synthetic vascular graft and other non-valvular cardiovascular devices||Any endoscopic procedure||Prevention of graft and device infection||Not recommended||1C+; no reported cases of infection associated with endoscopy|
|Prosthetic joints||Any endoscopic procedure||Prevention of septic arthritis||Not recommended||1C+; very low risk of infection|
Some societies (British Society of Gastroenterology) recommend antibiotic prophylaxis for patients with severe neutropenia (<0.5 × 10 9 /L) and/or advanced hematological malignancy who undergo endoscopic procedures that are known to be associated with a high risk of bacteremia.
Vascular grafts or prostheses
Antibiotic prophylaxis is not recommended for patients with vascular grafts or prostheses.
Patients with cirrhosis or portal hypertension
Patients with portal hypertension and ascites have an increased risk of bacteremia in the presence of an upper gastrointestinal bleed, while sclerotherapy is also associated with high rates of bacteremia. Patients with cirrhosis and bacterial infection have an increased risk of re-bleeding, and poor outcomes, while patients who receive antibiotic prophylaxis may have a reduced risk of variceal re-bleeding. A meta-analysis has shown that antibiotic prophylaxis in patients with variceal bleeding is associated with improved short-term survival.
Recommendation for patients with cirrhosis and upper GI bleeding:
All patients with cirrhosis who present with an upper gastrointestinal bleed should be given prophylactic antibiotics.
Antibiotic prophylaxis for specific endoscopic procedures
See Table 3 for recommended antibiotics based on the British Society of Gastroenterology guidelines.
|ERCP||Ciprofloxacin 750 mg PO 90 min pre-procedure|
|Gentamicin 1.5 mg/kg IV|
|OLT undergoing ERCP||Ciprofloxacin 750 mg 90 min pre-procedure|
|Gentamicin 1.5 mg/kg IV|
|Amoxicillin 1 g IV or vancomycin 20 mg/kg IV infused over at least one hour|
|EUS FNA cystic lesion||Co-amoxiclav 1.2 g IV|
|Ciprofloxacin 750 mg PO 90 min pre-procedure|
|3–5 day course of antibiotics post-procedure is usually given|
|Antibiotics should be given prior to performing EUS-FNA|
|PEG||Co-amoxiclav 1.2 g IV|
|Second or third generation cephalosporin (i.e. cefuroxime 750 mg IV)|
|Teicoplanin 400 mg IV can be used in patients who are penicillin allergic|
|Antibiotics should be given prior to commencing the procedure|
|Cirrhosis with upper-GI bleed||Piperacillin/tazobactam 4.5 g IV three times per day|
|Third generation cephalosporin (i.e. cefotaxime 2 g IV three times per day)|
Sepsis and cholangitis occur in 0.5–3% of ERCP procedures; however, a meta-analysis of five randomized, placebo-controlled trials failed to demonstrate a decrease in the incidence of cholangitis or sepsis with prophylactic antibiotic use. The exception to this rule is patients with incomplete biliary drainage, who have a high risk of sepsis. These patients should be given antibiotic coverage until adequate drainage is achieved. Patients with pancreatic pseudocysts are also at increased risk of infection, and should receive antibiotic cover.
Recommendations for patients undergoing ERCP:
Patients with biliary sepsis should be given appropriate antibiotics.
Patients with biliary obstruction with no evidence of infection do not require antibiotics, unless there is inadequate biliary drainage.
Patients with undrained biliary systems following ERCP (cholangiocarcinoma, primary sclerosing cholangitis with failed drainage procedure) are at high risk of ascending cholangitis and sepsis and should receive antibiotic cover until adequate drainage is achieved.
Patients with pancreatic pseudocysts are at increased risk of infection within the pseudocyst. Those patients undergoing an interventional endoscopic procedure (ERCP or EUS) should be given prophylactic antibiotics.
Some societies (BSG) recommend antibiotic prophylaxis for patients with an orthotopic liver transplant (OLT) who undergo an ERCP.
The risk of infection after EUS-FNA of a solid lesion is very low (0.4%). These patients do not require antibiotic prophylaxis. There is an increased risk of infection when samples are acquired from cystic lesions. Therefore, patients undergoing endoscopic fine-needle aspiration (EUS-FNA) of cysts should receive peri-procedural antibiotics. Most endoscopists also continue oral antibiotics for 3–5 days. There is insufficient evidence to recommend antibiotics following EUS-FNA of a solid lesion in the lower gastrointestional tract. The choice of whether or not to give antibiotics should be made on a case by case basis.
Recommendations for patients undergoing EUS:
Patients undergoing EUS-FNA of a cystic lesion, should have prophylactic antibiotics.
Prophylactic antibiotics are not required for EUS-FNA of solid lesions.
Patients undergoing pancreatic pseudocyst drainage should have prophylactic antibiotics.
A large number of studies, including two meta-analyses, have shown that antibiotic prophylaxis is effective at reducing wound infection rates using a single dose of an appropriate antibiotic. PEG insertion is associated with a high risk of peristomal wound infection and patients should be given antibiotic prophylaxis.
Recommendations for patients undergoing PEG:
Patients undergoing PEG insertion should receive antibiotic prophylaxis.
(Guidelines of the British Society of Gastroenterology)
(Guidelines of the American Society of Gastrointestinal Endoscopy)
(Guidelines from the American College of Cardiology and American Heart Association)
(Guidelines of the American Heart Association)