Antithrombotic drugs (anticoagulants, aspirin, and other antiplatelet agents) are used to treat cardiovascular disease and to prevent secondary thromboembolic events. These drugs are independently associated with an increased risk of gastrointestinal bleeding (GIB), and, when prescribed in combination, further increase the risk of adverse bleeding events. Clinical evidence to inform the choice of endoscopic hemostatic procedure, safe temporary drug cessation, and use of reversal agents is reviewed to optimize management following clinically significant GIB.
Key points
- •
Antithrombotic drugs are associated with a clinically significant risk of gastrointestinal bleeding.
- •
An important consideration is if endoscopic hemostasis (in itself) constitutes a high vs. low-risk procedure.
- •
A better understanding of the pharmacology, mechanism of action and clinical indications for common antiplatelet drugs is imperative for sound decision-making regarding drug cessation or continuation in the peri-endoscopic period.
- •
Management of anticoagulant associated bleeding in the emergent and urgent setting is still grounded in the principles of A (airway), B (breathing), and C (circulation).
- •
There is remarkably little data to inform the endoscopist’s decision of resumption of antithrombotic therapy.
Introduction
Current estimates of antithrombotic use in the United States are limited. The Reduction of Atherothrombosis for Continued Health (REACH) registry suggests that 70% of Americans (n = 25,686) are on acetylsalicylic acid (ASA) monotherapy; 13% are on ASA with a thienopyridine antiplatelet agent (ie, dual antiplatelet therapy [DAPT]), 8% are on anticoagulant or thienopyridine antiplatelet agent monotherapy, 4% are on ASA plus anticoagulant, and 1% are on thienopyridine agent plus anticoagulant or on all 3 antithrombotic agents concurrently. Data from the Department of Veterans’ Affairs (n = 78,133) show that 50.5% are on DAPT, 29.3% are on ASA plus anticoagulant, 13.8% are on anticoagulant plus thienopyridine antiplatelet agent, and 6.3% are on triple therapy with ASA plus anticoagulant plus thienopyridine agent. It is projected that, by 2030, greater than 40% of US adults (>25 million individuals) will have at least 1 form of cardiovascular disease, accompanied by an expected aggressive increase in antithrombotic drug use for prevention of myocardial infarction (MI), stroke (cardiovascular accident [CVA]), and thromboembolic disorders (deep venous thromboembolism or pulmonary embolism) in patients who have already had a prior event (ie, for secondary cardioprophylaxis).
These drugs are associated with an important and clinically relevant gastrointestinal bleeding (GIB) risk. Abraham and colleagues showed the magnitude of risk associated with the use of antithrombotic drugs used in dual and triple combinations. The 1-year number needed to harm for common dual therapy strategies (ASA plus thienopyridine agent, ASA plus anticoagulant, or anticoagulant plus thienopyridine agent) as well as triple therapy (ASA plus thienopyridine agent plus anticoagulant) is less than 93 patients to incur 1 additional upper gastrointestinal (GI) bleed, less than 23 to incur 1 additional lower GI bleed, less than 51 to incur 1 additional blood transfusion, and less than 67 patients to incur 1 additional GI bleed–related hospitalization.
These estimates may represent just the “tip of the iceberg” because they fail to include the impact of GIB associated with the new oral anticoagulants, which are known to increase the risk of GIB 3-fold when combined with ASA and a thienopyridine agent. Furthermore, with the aging US population, GIB is likely to increase because of the presence of multiple concomitant risk factors in this population: (1) advancing age, (2) multiple medical comorbidities, and (3) increased use of antiplatelet and anticoagulant agents in combination. The synergism of these risk factors is likely to change the epidemiology of GIB in North America.
This article focuses on the management of antithrombotic agents in the periendoscopic period surrounding an acute, clinically significant GIB, requiring endoscopic intervention. These patients include those with hemodynamic compromise, greater than or equal to a 2-g reduction in hemoglobin, or overt signs of GIB (melena, hematemesis, coffee-ground emesis, and hematochezia).
This article addresses the following clinical questions:
- 1.
Is endoscopic hemostasis considered a high-risk or low-risk procedure?
- 2.
How should antiplatelets be managed when the patient is bleeding?
- 3.
How should anticoagulants be managed when the patient is bleeding?
- a.
How should the novel oral anticoagulants (NOACs) be managed in the urgent setting?
- b.
What are the new target-specific NOAC reversal agents?
- a.
- 4.
Should the patient be bridged if stopping anticoagulation?
- 5.
When should antithrombotics be restarted?
Is Endoscopic Hemostasis Considered a High-risk or Low-risk Procedure?
An important consideration is whether endoscopic hemostasis (in itself) constitutes a high-risk versus low-risk procedure. The American Society of Gastrointestinal Endoscopy considers a low-risk procedure to be a procedure that is associated with a clinical rate of bleeding of 1.5% or less, in the absence of antithrombotic therapy. If a procedure with a risk greater than 1.5% is considered high risk, many of the commonly performed hemostatic procedures would be in this category. Some procedures (such as hemostatic clip placement, injection) remain ill-defined in terms of postprocedural bleeding risk ( Table 1 ). However, few endoscopic procedures are associated with closed-space bleeding (ie, retroperitoneal, intrathoracic, or pericardial), unless a major perforation is incurred.
| Endoscopic Procedure | Low-risk Bleeding (<1.5%) | High-risk Bleeding (>1.5%) |
|---|---|---|
| Diagnostic EGD or colonoscopy (with or without biopsy) | X | — |
| Nonthermal removal of small polyps (<1 cm) | X | — |
| Coagulation or ablation of tumors or vascular lesions (includes APC, bipolar cautery, and laser ablation) | — | X |
| Large (>1 cm) polypectomy | — | X |
| Variceal band ligation | — | X |
| Hemostatic clip placement | X (unknown risk) | — |
| Injection therapy | X (unknown risk) | — |
| Bipolar cautery | — | X |
Consequences of the procedurally-induced bleed also need to be considered. Baron and colleagues classified the severity of bleeding consequences based on expert consensus from a single institution and extrapolation from existing guidelines and consensus statements regarding endoscopic risk. Mild consequences of postprocedural bleeding would include incomplete or aborted procedures, need for repeat endoscopy, transfusion or interventional radiology, the need for unplanned ventilation or anesthesia support, postprocedural medical consultation, or an unplanned hospital stay of less than 3 nights or intensive care admission of less than 1 night. Moderate consequences include an unplanned admission for 4 to 10 nights related to the procedural bleeding, intensive care unit admissions greater than 1 night, need for surgery, or permanent disability. Major consequences include admissions for greater than 10 nights, intensive care unit admissions greater than 1 night, and death.
What remains less clear is the underlying risk of performing endoscopic hemostasis in a patient in whom platelet dysfunction is expected because of the use of a pharmacologic agent, and in whom prompt resumption of the antithrombotic agent is necessary to prevent an adverse thromboembolic event. Endoscopic hemostasis should be considered an activity that carries with it a high risk of postprocedural bleeding. The magnitude of risk is at least 1.5%, and possibly much higher, based on the patient’s underlying antithrombotic regimen; preexisting nonpharmacologic coagulopathy and associated comorbidities, including renal and/or hepatic dysfunction; or the presence of carcinoma.
- •
Performing endoscopic procedures on patients currently prescribed antithrombotic regimens is both warranted and safe, providing clinicians consider the thrombotic risk of temporary interruption of drugs, and the relative risk of endoscopic maneuvers on subsequent bleeding, and promptly restart regimens when hemostasis is assured or provide temporary bridge therapy in patients in whom hemostasis is uncertain.
How Should Antiplatelets Be Managed When the Patient Is Bleeding?
A better understanding of the pharmacology, mechanism of action, and clinical indications for common antiplatelet drugs is imperative for sound decision-making regarding drug cessation or continuation in the periendoscopic period. Commonly used antiplatelet agents include ASA; dipyridamole; and the thienopyridine drugs clopidogrel, prasugrel, and ticagrelor. ASA inhibits both cyclooxygenase (COX)-1 and COX-2 and causes irreversible inhibition of platelet function. Dipyridamole inhibits thrombus formation by inhibition of the phosphodiesterase enzymes that break down cAMP and cGMP, impairing platelet function and promoting arteriolar smooth muscle relaxation. The time required to recover adequate platelet function after ASA and dipyridamole use is ∼7 to 10 days.
The thienopyridine agents, clopidogrel, prasugrel, and ticagrelor, inhibit the P2Y12 receptor on the platelet to inhibit platelet aggregation. Inhibition is irreversible for clopidogrel and prasugrel and reversible for ticagrelor. The antiplatelet effect can last between 3 and 9 days depending on the agent. The newer antiplatelet agents vorapaxar and atopaxar, inhibit the protease-activated receptor-1 on the platelet. These drugs are less commonly used because the increased risk of serious bleeding (especially intracranial bleeding) outweighs the modest finding of efficacy in randomized controlled trials (RCTs) among patients with acute coronary syndrome, thus, they are not discussed further in this article.
Aspirin monotherapy
ASA is a COX inhibitor that is used alone or in combination with other antithrombotic therapies. It is the cornerstone of cardiac prevention strategies for patients at greater than 10% 5-year risk of heart attack or stroke (ie, primary cardioprophylaxis) and for secondary prevention of cardioembolic events in patients who have had a prior acute coronary syndrome event or stroke.
Acetylsalicylic acid plus dipyridamole
Dipyridamole is commonly used in the United States as a combination pill with ASA (Aggrenox) for the secondary prevention of stroke and transient ischemic attack. It is not approved as monotherapy for stroke prevention, and the use of ASA plus dipyridamole with the thienopyridine agent, clopidogrel, has generally been abandoned because of the high risk of bleeding adverse events. Discontinuation of ASA plus dipyridamole for 7 to 10 days returns platelet function to normal, in the absence of other coagulopathies.
Dual antiplatelet therapy (acetylsalicylic acid plus thienopyridine agent)
Thienopyridine agents include clopidogrel [Plavix], prasugrel [Effient], ticlopidine [Ticlid], and ticagrelor [Brilinta]. Current national cardiology guidelines include prescription of clopidogrel, ticagrelor, and prasugrel plus ASA following acute coronary syndrome for up to 12 months following unstable angina or non–ST elevation MI managed without percutaneous coronary intervention (PCI) and for at least 14 days (12 months in some patients) following ST segment elevation myocardial infarction. Following a stent insertion, ASA must be continued indefinitely and clopidogrel or ticagrelor prescribed for up to 12 months following bare metal stent insertion and at least 12 months following drug-eluting stent placement.
- •
When endoscopists are considering altering dual antiplatelet therapy (DAPT) with ASA plus thienopyridine agent, it is important to remember that thromboembolic risk depends on 3 factors:
- ○
The indication for the antiplatelet therapy
- ○
The consequences of thromboembolic event
- ○
The presence of additional thromboembolic risk factors
- ○
Low versus high thromboembolic risk conditions
Low thromboembolic risk conditions include uncomplicated or paroxysmal nonvalvular atrial fibrillation, bioprosthetic valves, a mechanical valve in the aortic position, and deep vein thrombosis. High-risk thromboembolic conditions include atrial fibrillation associated with one of the following: valvular heart disease and the presence of prosthetic valves, an ejection fraction less than 35%, hypertension, diabetes mellitus, age greater than 75 years, and a history of thromboembolic event. Additional high thromboembolic risk conditions include a mechanical valve in any position and previous thromboembolic event, prior stent occlusion, a recently placed coronary stent (∼1 year), acute coronary syndrome, and PCI after MI.
- •
Periods of time when thromboembolic risk are the highest in patients with coronary stents:
- ○
First 90 days following acute coronary syndrome
- ○
First 30 to 45 days after PCI and bare metal stent insertion
- ○
First 3 to 6 months following PCI and drug-eluting stent insertion
- ○
Following acute coronary syndrome, regardless of whether the patient has been medically treated or undergone PCI, the most dangerous period of time to alter DAPT is in the first 90 days following the event. In this time the risk is 2-fold higher for cardiac death or MI with clopidogrel discontinuation. It is also important to remember that, after PCI and bare metal stent insertion, the highest risk of stent occlusion is in the first 30 to 45 days, and within the first 6 months of drug-eluting stent placement, so this is not the time to alter DAPT. One in 5 patients who experience a first definite stent thrombosis experience a second stent thrombosis and that risk can remain at ∼2.9% over the next 3 years.
- •
For an endoscopic procedure with a high risk of GI bleeding, the thienopyridine can be discontinued for a short period of time (∼5–7 days) as long as the ASA is continued.
There is no increased risk of postprocedural bleeding associated with continued use of ASA and, in high-risk cardiac patients, discontinuation can increase 30-day mortality.
- •
Premature and complete discontinuation of antithrombotic therapy in anticipation of an endoscopic procedure can result in stent occlusion, MI, and mortality in 50% of patients.
- •
Elective diagnostic endoscopy can safely be performed without cessation of DAPT.
Glycoprotein IIb/IIIa receptor inhibitors
The intravenous agents abciximab (ReoPro), eptifibatide (Integrilin), and tirofiban (Aggrastat) are used following acute coronary syndrome in patients undergoing PCI or endovascular interventions. They prevent fibrinogen-mediated platelet aggregation, thrombus formation, and distal thromboembolism. Associated adverse events include significant rates of major bleeding and, with some, thrombocytopenia that further exacerbate bleeding events. Duration of effect ranges from 1 to 2 seconds (tirofiban) to 4 to 24 hours, dictating their rate of intravenous infusion.
- •
In situations of major GI hemorrhage, aggressive volume resuscitation, the use of inotropes if necessary, transient discontinuation of the glycoprotein IIb/IIIa infusion, and platelet transfusion can be helpful to promote hemostasis in anticipation of endoscopic or surgical hemostatic interventions. In rare cases, hemodialysis may be required (tirofiban)
How Should Anticoagulants Be Managed When the Patient Is Bleeding?
Inhibition of single or multiple steps in the coagulation cascade is the mechanism of action associated with anticoagulants.
Heparin derivatives
Unfractionated heparin is administered parenterally and has a short half-life of 60 to 90 minutes. Complete dissipation of anticoagulant effect occurs after 3 to 4 hours. The risk of bleeding associated with parenteral unfractionated heparin is less than 3% in clinical trials (among patients with deep venous thrombosis [DVT]); a risk that increases with dose escalation and age greater than 70 years. Low-molecular-weight heparin agents such as enoxaparin and dalteparin are administered subcutaneously and are frequently used for bridging therapy in patients during temporary interruption of oral anticoagulants. These agents are also used therapeutically in the treatment of DVT and have a lower risk of major bleeding than unfractionated heparin.
Fondaparinux is specifically approved for perioperative thromboembolic prophylaxis and for the initial therapy for both DVT and pulmonary embolism. What differentiates this agent from others in this drug class is its high affinity for antithrombin III, which potentiates the inhibitory effect of factor Xa. Before a high-risk endoscopic procedure (eg, endoscopic mucosal resection, endoscopic ultrasonography with fine-needle aspiration, endoscopic submucosal dissection, variceal banding), a minimum of 36 hours of drug interruption is required.
Warfarin
For more than 80 years, warfarin was the only clinical choice for oral anticoagulation. This anticoagulant inhibits the vitamin K–dependent clotting factors II, VII, IX, and X, as well as proteins C and S. The risk of adverse bleeding events, unpredictable pharmacodynamic response, significant potential for drug-drug interactions, and delayed onset of action were drawbacks of this agent. The necessity for frequent monitoring, dose adjustment, and compliance with dietary restrictions has limited its popularity with patients. The anticoagulant effect can be predictably reduced following temporary interruption; the International Normalized Ratio (INR) decreases to less than 1.5 in 93% of patients within 5 days. This fact coupled with well-established algorithms for bridging therapy using low-molecular-weight heparin or unfractionated heparin products during periods of temporary interruption makes it a popular choice for clinical use.
The risk of warfarin-associated bleeding is determined by the intensity of the anticoagulant effect, baseline patient characteristics, and duration of therapy. A targeted INR of 2.5 (range, 2.0–3.0) is associated with a lower risk of bleeding than an INR greater than 3.0. Reversal of warfarin-induced anticoagulant effect can be achieved by transfusion of hemostatic blood products, such as fresh frozen plasma, and provision of vitamin K. Normalization of the INR is unnecessary and does not reduce rebleeding risk, but does contribute to significant delays in endoscopy, which delays discovery of important endoscopic stigmata in 83% of cases. Hemostatic therapy (injection therapy, heater probe, hemoclips) is very effective even with a moderately increased INR (up to 2.7).
- •
Endoscopic hemostasis can safely and reliably be performed in anticoagulated patients with INRs up to 2.7.
- •
No need to normalize the INR before proceeding with endoscopic therapy.
- •
Rebleeding rates are similar with and without anticoagulant reversal.
Novel oral anticoagulants
NOACs include direct thrombin inhibitor, dabigatran etexilate (Pradaxa), and direct oral factor Xa inhibitors, rivaroxaban (Xarelto), apixaban (Eliquis), and edoxaban (Lixiana). These agents, developed to overcome the limitations of warfarin with their rapid and predictable pharmacodynamic response and fixed once or twice daily dosing regimens, have quickly become popular with physicians and patients alike. Pivotal cardiac clinical trials have shown an unexpected increase in the risk of GI bleeding based on specific agents and indications for use.
Apixaban seems to have the lowest risk of GI bleeding (hazard ratio [HR], 0.76; 95% confidence interval [CI], 0.98–1.15). A 46% increase in GI bleeding (HR, 1.46; 95% CI, 1.19–1.78) is observed with rivaroxaban. Dabigatran, at 150 mg twice a day, is associated with a 50% increase in bleeding risk (HR, 1.50; 95% CI, 1.1.9–1.89), especially in the elderly. When combined with DAPT, a 3-fold increased risk of major bleeding is observed (HR, 3.03; 95% CI, 2.20–4.16) with a number needed to harm of 111 to generate 1 additional major bleed (which includes GI bleeds, intracranial hemorrhage, clinically overt signs of hemorrhage associated with a reduction in hemoglobin level ≥5 g/dL, and fatal bleeding that results in death within 7 days). Lamberts and colleagues report fatal NOAC-related intracranial hemorrhage in 48%, and fatal GI bleeds in 45.3%. Nonfatal GI bleeds are also common (33.8%) and represent the most common NOAC-related bleeding complication.
The large NOAC RCTs did not provide much guidance for periprocedural management of these drugs. A post-hoc analysis of the RE-LY (Randomized Evaluation of Long-term Anticoagulation Therapy) trial (dabigatran) identified 3033 patients who underwent surgery or invasive procedures during the observation period of the parent trial. The mean time of drug discontinuation before the procedure was 49 hours (range, 38–85 hours) and the observed rate of postprocedural bleeding was equivalent to that seen in the warfarin control group: 6.5% after elective major surgery and 17.7% after emergency surgery. The most aggressive mucosal disruption caused by an endoscopic procedure would not be equivalent to a major elective surgical wound, so the predicted rate of bleeding following an endoscopic procedure is likely to be between 1.5% and 6.5%. Although the overall risk of postprocedural bleeding is similar to that of warfarin, a shorter period of drug cessation is required (with normal renal function); 2 days with dabigatran versus 5 days with warfarin.
Based on limited clinical data and pharmacodynamic studies of these agents, recommendations for NOAC management in the elective periendoscopic period have been proposed ( Table 2 ). These data take advantage of the unique properties of the NOACs, which require less than 4 hours to achieve maximum effect, have a short half-life (<15 hours), and are excreted primarily through the kidneys. Timing of discontinuation of the drug before endoscopy is based on the patient’s creatinine clearance (CrCl) and the anticipated bleeding risk of the endoscopic procedure.
Related posts:
Advances and Improvements in the Management of Upper Gastrointestinal Bleeding
Epidemiology and Risk Factors for Upper Gastrointestinal Bleeding
Endoscopic Diagnosis and Therapy in Gastroesophageal Variceal Bleeding
Injection and Cautery Methods for Nonvariceal Bleeding Control
Unusual Causes of Upper Gastrointestinal Bleeding
Tips and Tricks on How to Optimally Manage Patients with Upper Gastrointestinal Bleeding
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
