American College of Gastroenterology Canadian Association of Gastroenterology

Journal of the Canadian Association of Gastroenterology, 2022, XX, 1–18
https://doi.org/10.1093/jcag/gwac010
Advance access publication 17 March 2022
Clinical Guidelines
American College of Gastroenterology-Canadian

Association of Gastroenterology Clinical Practice
Guideline: Management of Anticoagulants and
Antiplatelets During Acute Gastrointestinal Bleeding and
Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

the Periendoscopic Period
Neena S. Abraham, MD, MSc (Epi), FACG1, Alan N. Barkun, MD, MSc (Epi), FACG, CAGF2,
Bryan G. Sauer, MD, MSc (Clin Res), FACG3, James Douketis, MD4, Loren Laine, MD, FACG5,6,
Peter A. Noseworthy, MD7, Jennifer J. Telford, MD, MPH, FACG, CAGF8, and
Grigorios I. Leontiadis, MD, PhD, CAGF9
1
Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Scottsdale, Arizona, USA;
2
Division of Gastroenterology, Department of Medicine, McGill University, Montreal, Quebec, Canada;
3
Division of Gastroenterology and Hepatology, University of Virginia, Charlottesville, Virginia, USA;
4
Department of Medicine, St. Joseph's Healthcare Hamilton and McMaster University, Hamilton, Ontario, Canada;
5
Yale School of Medicine, New Haven, Connecticut, USA;
6
Virginia Connecticut Healthcare System, West Haven, Connecticut, USA;
7
Department of Cardiovascular Diseases, Electrophysiology, Mayo Clinic, Rochester, Minnesota, USA;
8
Division of Gastroenterology, Department of Medicine, St. Paul's Hospital, University of British Columbia, Vancouver, British Columbia, Canada;
9
Division of Gastroenterology and Farncombe Family Digestive Health Research Institute, Department of Medicine, McMaster University,
Hamilton, Ontario, Canada.
This article is being published jointly in The American Journal of Gastroenterology and the Journal of Canadian Association of Gastroenterolgy.
SUPPLEMENTARY MATERIAL accompanies this paper at http://links.lww.com/AJG/C416, http://links.lww.com/AJG/C417, and http://links.lww.com/AJG/
C418.
Correspondence: Neena S. Abraham, MD, MSc (Epi), FACG. E-mail: [email protected].
Abstract
We conducted systematic reviews of predefined clinical questions and used the Grading of Recommendations, Assessment, Development
and Evaluations approach to develop recommendations for the periendoscopic management of anticoagulant and antiplatelet drugs during
acute gastrointestinal (GI) bleeding and the elective endoscopic setting. The following recommendations target patients presenting with acute
GI bleeding: For patients on warfarin, we suggest against giving fresh frozen plasma or vitamin K; if needed, we suggest prothrombin complex
concentrate (PCC) compared with fresh frozen plasma administration; for patients on direct oral anticoagulants (DOACs), we suggest against
PCC administration; if on dabigatran, we suggest against the administration of idarucizumab, and if on rivaroxaban or apixaban, we suggest
against andexanet alfa administration; for patients on antiplatelet agents, we suggest against platelet transfusions; and for patients on cardiac
acetylsalicylic acid (ASA) for secondary prevention, we suggest against holding it, but if the ASA has been interrupted, we suggest resump-
tion on the day hemostasis is endoscopically confirmed. The following recommendations target patients in the elective (planned) endoscopy
setting: For patients on warfarin, we suggest continuation as opposed to temporary interruption (1–7 days), but if it is held for procedures with
high risk of GI bleeding, we suggest against bridging anticoagulation unless the patient has a mechanical heart valve; for patients on DOACs,
we suggest temporarily interrupting rather than continuing these; for patients on dual antiplatelet therapy for secondary prevention, we sug-
gest temporary interruption of the P2Y12 receptor inhibitor while continuing ASA; and if on cardiac ASA monotherapy for secondary prevention,
we suggest against its interruption. Evidence was insufficient in the following settings to permit recommendations. With acute GI bleeding in
patients on warfarin, we could not recommend for or against PCC administration when compared with placebo. In the elective periprocedural
endoscopy setting, we could not recommend for or against temporary interruption of the P2Y12 receptor inhibitor for patients on a single P2Y12
inhibiting agent. We were also unable to make a recommendation regarding same-day resumption of the drug vs 1–7 days after the procedure
among patients prescribed anticoagulants (warfarin or DOACs) or P2Y12 receptor inhibitor drugs because of insufficient evidence.
INTRODUCTION antiplatelet drugs such as the P2Y12 receptor inhibitors

Antithrombotic drugs including vitamin K antagonists (VKAs; (clopidogrel, prasugrel, and ticagrelor), and acetylsalicylic
warfarin and acenocoumarol), direct oral anticoagulants acid (ASA) are used in the management of patients with atrial
(DOACs; apixaban, dabigatran, edoxaban, and rivaroxaban), fibrillation, ischemic heart disease, venous thromboembolism,
© The Authors, 2022. This article is a co-publication between the Journal of the Canadian Association of Gastroenterology and The American Journal of
Gastroenterology
2 Journal of the Canadian Association of Gastroenterology, 2022, Vol. XX, No. XX
and valvular heart disease. These drugs also increase the risk gastroenterologists who served as the GRADE methodologists
of gastrointestinal (GI) bleeding from luminal sources such as (G.I.L. and B.S.). No patients were included in the guideline
ulcers or diverticula and after endoscopic procedures (1–3). process. The panel developed, prioritized, and finalized the
Standardized, evidence-based protocols are lacking to in- clinical questions in Population, Intervention, Comparator,
form best practices before and after endoscopic procedures and Outcome (PICO) format through teleconferences be-
in urgent and elective settings. Furthermore, uncertainty re- fore systematic literature reviews. The critical outcomes were
garding best practice recommendations and associated levels 7-day further bleeding and 30-day thrombotic events for pa-
of evidence has led to significant variation in adherence to tients with acute GI bleeding and 30-day bleeding and 30-day
guideline-directed practices (4). thrombotic events after elective endoscopic procedures. The
The American College of Gastroenterology (ACG) and final PICO questions were shared with the leadership of the
the Canadian Association of Gastroenterology (CAG) con- ACG Practice Parameters Committee and the CAG Clinical
vened an international, multisociety, and multidisciplinary Affairs Committee.

working group to create a focused, pragmatic guideline The editorial office of the Cochrane Gut Group at
after distillation of published literature to inform clinical McMaster University developed and ran searches in
practice in the periendoscopic period. In keeping with the MEDLINE, EMBASE, and CENTRAL for randomized con-
Grading of Recommendations, Assessment, Development trolled trials (RCTs), controlled or uncontrolled observational
and Evaluations (GRADE) approach (5), the most pertinent studies, and systematic reviews of any study design published
clinical questions guided the systematic review of the litera- in the English language as full text (conference abstracts were
ture, with the resulting rigorous methodological evaluation not included) between January 1, 1995 (January 1, 1985, for
of the available published data informing recommenda- some searches), and August 13, 2020. Full details of search
tions. In this document, we propose an evidence-based ap- strategies can be found in Supplementary Digital Content (see
proach to periprocedural antithrombotic drug management Appendix 1, http://links.lww.com/AJG/C416). Each identified
in common emergent and elective settings addressing clin- abstract was screened for eligibility in duplicate by at least 2
ical questions related to (i) temporary interruption of anti- of the 4 voting gastroenterologists. Potentially eligible studies
coagulant and antiplatelet drugs; (ii) reversal of anticoagulant were assessed as full-text articles by the GRADE method-
and antiplatelet drugs; (iii) periprocedural heparin bridging; ologists or 1 of the 4 voting gastroenterologists. A GRADE
and (iv) postprocedural resumption of anticoagulant and methodologist verified data extraction. An evidence map was
antiplatelet drugs. prepared for each PICO question. The panel reviewed the pre-
This document does not cover all possible clinical situ- liminary evidence map, proposed additional articles, and as-
ations where multidisciplinary guidance may be necessary to sisted in supplementary literature searches targeting broader
manage periendoscopic antithrombotic therapy. Nor does it populations when gaps in the evidence were identified. Where
address the rapidly evolving menu of endoscopic approaches appropriate, more recent publications available after the
developed to minimize intraprocedural and postprocedural formal literature search and evidentiary review are discussed
bleeding in situations such as removing large colonic polyps for contextual information if deemed to provide critical add-
(6). Because of insufficient evidence, the panel could not rec- itional contemporary insight.
ommend a best practice for all clinical questions. These clin- The 2 GRADE methodologists prepared assessments of the
ical situations are identified as priorities for future research. risk of bias of each included study and developed complete
evidence reports, including a summary of evidence tables (see
Appendix 2, Supplementary Digital Content, http://links.lww.
METHODS com/AJG/C417). The certainty of the evidence for each PICO
These guidelines are established to support clinical practice question was categorized as very low, low, moderate, or high
and suggest preferable approaches to a typical patient with depending on the assessment of (i) limitations in the design
a particular medical problem based on the currently avail- and execution of the studies, (ii) indirectness, (iii) inconsist-
able published literature. When exercising clinical judgment, ency, (iv) imprecision, and (v) other considerations including
particularly when treatments pose significant risks, healthcare publication bias, according to the GRADE approach (7,8).
providers should incorporate this guideline in addition to Manuscripts initially deemed potentially eligible but eventu-
patient-specific medical comorbidities, health status, and ally excluded are listed in Supplementary Digital Content (see
preferences to arrive at a patient-centered care approach. Appendix 3, http://links.lww.com/AJG/C418) with reasons
The methods for this guideline were agreed on a priori by for exclusion. Each GRADE methodologist, in turn, prepared
the ACG and the CAG with the express intent to codevelop half of the evidence reports, whereas the other methodologist
high-quality multisociety guidelines that reduce duplication double-checked them, providing feedback until agreement
of effort and improve impact. The methods have followed the was achieved. For each PICO, 3 versions of the wording of the
GRADE approach (5). The target population of this guide- potential recommendation were prepared a priori (in favor,
line is patients receiving anticoagulants or antiplatelet drugs against, or unable to recommend). The opinions of individual
who are (i) hospitalized or under observation with acute GI content experts were sought for specific issues. The evidence
bleeding or (ii) undergoing inpatient or outpatient elective GI reports and risk of bias tables were shared with the whole
endoscopic procedures. The target audience for this guideline panel on April 16, 2021, and discussed by email. The final-
includes healthcare providers, public health policymakers, pa- ized document was shared before the voting videoconference
tients, and caregivers. meetings on May 8 and 15, 2021.
The guideline panel was led by 2 gastroenterology cochairs One cochair (A.N.B.) and 1 GRADE methodologist (G.I.L.)
(N.S.A. and A.N.B.). It included 6 voting content experts—4 moderated the voting videoconference meetings. For each
gastroenterologists (N.S.A., A.N.B., L.L., and J.T.), 1 cardi- PICO, the GRADE methodologist presented a summary of
ologist (P.A.N.), 1 thrombosis expert (J.D.), and 2 nonvoting the evidence, including the direction and magnitude of effect
Journal of the Canadian Association of Gastroenterology, 2022, Vol. XX, No. XX 3
for desirable and undesirable outcomes and the certainty of GUIDELINE STATEMENTS
the evidence. After which, the panel discussed results. All do-
Management of antithrombotic agents in the
mains of the Evidence-to-Decision Framework (9), including
setting of acute GI bleeding
the certainty of evidence on the balance between desirable
and undesirable outcomes, evidence and assumptions about The first 10 guideline statements address the management
patient values and preferences, feasibility, acceptability, and of antithrombotic agents in the setting of acute GI bleeding.
resource use associated with alternative management options, Acute GI bleeding is defined as patients hospitalized or under
were reviewed, agreed on, summarized, and tabulated in real observation with acute overt GI bleeding (upper and/or lower)
time for the PICO question being assessed (7,9). Notes were manifesting as melena, hematochezia, or hematemesis. Life-
taken with regards to qualifiers and dissenting opinions. The threatening hemorrhage is defined as major clinically overt or
6 voting panel members then voted on the direction of the apparent bleeding, resulting in hypovolemic shock or severe
recommendation (in favor vs against) for that PICO question hypotension requiring pressors or surgery; or associated with

with its corresponding wording. The predetermined threshold a decrease in hemoglobin of >5 g/dL, or requiring transfusion
vote for consensus was 75% (i.e., 5 of 6 panel members). If of ≥ 5 units of packed red blood cells, or causing death (11).
consensus was not reached, the topic was further discussed,
VKA reversal
and reasons for disagreement were sought, with the panel
voting for a second time. If the 75% threshold could still not
be reached, the conclusion that “we could not reach a recom- 1. For patients on warfarin who are hospitalized or under
mendation for or against” the intervention was assigned to observation with acute GI bleeding, we suggest against
that PICO question. fresh frozen plasma (FFP) administration (conditional
If the 75% threshold was reached, provided the certainty of recommendation, very low certainty of evidence).
the evidence was moderate or high, panel members intended
to discuss and vote on the strength of recommendation (strong Summary of evidence. For this recommendation, no eligible
vs conditional). If 75% of the members voted for strong, the studies specifically addressing patients with GI bleeding were
recommendation would begin with “we recommend that identified by literature searches. The observational studies
….” Strong recommendations imply that most informed pa- identified were cohort studies without a comparator arm, or
tients would choose the recommended course of action, and the study did not report separate results for clinical outcomes
clinicians should provide it to most patients (7). If less than in patients with GI bleeding. It is thus not possible to infer with
75% of the members voted for strong, the recommendation any certainty whether administering FFP can benefit, harm, or
would be considered conditional and began with the words make no difference in these patients compared with no reversal.
“we suggest that….” Conditional recommendations indicate Pertinent studies included a small cohort of 41 warfarin-
that most individuals in this situation would want the sug- treated patients requiring rapid reversal (12), with 12 re-
gested course of action. Still, others would not, and clinicians ceiving FFP, 29 receiving clotting factor concentrates, and all
should help each patient make decisions consistent with their receiving vitamin K 1–5 mg intravenously. No clinical out-
risks, values, and preferences, ideally using decision aids. comes were measured, but in the 12 patients given FFP, the
Recommendations with low or very low certainty of evidence international normalized ratio (INR) did not normalize (range
were designated as conditional by default (without voting on 1.6–3.8, mean 2.3), indicating an ongoing anticoagulated
the strength), although such recommendations could have state in all patients. In a case-control study of 267 patients
still been considered as strong if they had fulfilled criteria for with major bleeding prescribed VKA for venous thrombo-
1 of the 4 “paradigmatic situations” (10). A search of con- embolism, 78 patients had GI bleeding, but no results were
temporary studies and recent systematic reviews was also reported for the GI bleeding outcomes (13). In a multivariable
performed and detailed in the evidence profile to inform the analysis that failed to adjust sufficiently for confounding, FFP
panel deliberations concerning the preferences of providers use was associated with a higher risk of thrombotic events
and patients for a cardiovascular event vs a GI bleeding event (OR: 4.22; 95% CI: 1.25–14.3) (13).
(see Appendix 2, Supplementary Digital Content, http://links. Three additional RCTs which lacked the comparator of
lww.com/AJG/C417 pages 3–6). interest (i.e., placebo) provide cohort-type data that further
Each voting panel member, including the 2 cochairs, pre- inform this recommendation. Sarode et al. randomized 202
pared a draft for designated sections after the voting video- patients on a VKA with an INR ≥2.0 and major bleeding to
conference meeting. The 2 cochairs subsequently edited and FFP (n = 104) vs 4-factor prothrombin complex concentrate
merged these into a single manuscript. The final version (PCC), while both arms received vitamin K (5–10 mg intra-
was reviewed and approved unanimously. The final manu- venously). In the FFP arm, 58 patients had GI bleeding with
script was peer-reviewed by the ACG Practice Parameters excellent or good hemostatic efficacy achieved in 75.9%
Committee, CAG Clinical Affairs Committee, the ACG (14). Additional outcome measures were reported only for
Board of Trustees, the CAG chair of Clinical Practice, the all-cause bleeding and included thrombotic events in 7 of
CAG vice president for Clinical Affairs, the CAG Board of 109, mortality in 6 of 103, rapid INR reduction in 10 of 109,
Directors, and the CAG membership at large (to whom the and fluid overload after 10–45 days in 14 of 109. Smaller
document was made available for 2 weeks). For each PICO RCTs by Steiner et al. (15) (N = 50, 23 in the FFP arm) and
question, the evidence table that summarizes the data and Boulis et al. (16) (N = 13, 8 in the FFP arm) assessed patients
the grading of that evidence is in Supplementary Digital with intracranial hemorrhage, reporting thromboembolic
Content (see Appendix 2, http://links.lww.com/AJG/C417). events in 2 of 23 and 1 of 8 patients, respectively, and an
A complete list of guideline statements, the strength of rec- INR ≤2 within 3 hours of treatment in 2 of 23 patients; sig-
ommendation, and the certainty of the evidence is found in nificant complications from fluid overload were noted in 5
Tables 1 and 2. of 8 patients.
Table 1. Guideline statements, the strength of recommendation, and certainty of the evidence for the management of antithrombotic agents in the
setting of acute GI bleed
Management of antithrombotic agents in the setting of acute GI bleed
Vitamin K antagonist reversal

1. For patients on warfarin who are hospitalized or under observation with acute GI bleeding, we suggest against FFP administration (conditional
recommendation, very low certainty of evidence).
2. For patients on warfarin who are hospitalized or under observation with acute GIB, we could not reach a recommendation for or against PCC
administration.
3. For patients on warfarin who are hospitalized or under observation with acute GIB, we suggest PCC administration compared with FFP ad-
ministration (conditional recommendation, very low certainty of evidence).

4. For patients on warfarin who are hospitalized or under observation with acute GIB (upper and/or lower), we suggest against the use of vitamin
K (conditional recommendation, very low certainty of evidence).
Direct thrombin inhibitor reversal (dabigatran)
5. For patients on dabigatran who are hospitalized or under observation with acute GIB, we suggest against the administration of idarucizumab
(conditional recommendation, very low certainty of evidence).
Reversal of rivaroxaban/apixaban with andexanet alfa
6. For patients on rivaroxaban or apixaban who are hospitalized or under observation with acute GIB, we suggest against andexanet alfa admin-
istration (conditional recommendation, very low certainty of evidence).
Reversal of direct oral anticoagulant with PCC
7. For patients on DOACs who are hospitalized or under observation with acute GIB, we suggest against PCC administration (conditional recom-
mendation, very low certainty of evidence).
Reversal of antiplatelet with platelet transfusion
8. For patients on antiplatelet agents who are hospitalized or under observation with acute GIB, we suggest against platelet transfusions (condi-
tional recommendation, very low certainty of evidence).
Holding ASA vs continuing ASA
9. For patientswithGIbleeding oncardiacASAfor secondary prevention,we suggest against holding the ASA (conditional recommendation, very
low certainty of evidence).
Resumption of ASA after endoscopic hemostasis
10. For patients with GI bleeding on ASA for secondary cardiovascular prevention whose ASA was held, we suggest the ASA be resumed on the
day hemostasis is endoscopically confirmed (conditional recommendation, very low certainty of evidence).
ASA, acetylsalicylic acid; FFP, fresh frozen plasma; DOAC, direct oral anticoagulant; GI, gastrointestinal; GIB, GI bleeding; PCC, prothrombin complex
concentrate.
Conclusions. Although there is biological plausibility of FFP and results on the indirect outcome of INR reversal (17). This
administration to reverse VKA in patients with GI bleeding, study was regarded as noncomparative data with the inclusion
there exists only very low certainty evidence, given serious of the PCC group only.
concerns of risk of bias, imprecision, and indirectness. The panel From 7 studies, there were 223 patients on warfarin, all ex-
also considered the low cost of FFP, relevant patient utilities, periencing major bleeding and treated with PCC (14,15,17–
and the potential increased risk of transmission of infectious 21). Of these, 38.6% had GI bleeding. All patients received
agents with FFP administration. The panel suggested that 4-factor PCC at various doses, with vitamin K administered to
FFP should not be used routinely but could be considered for most patients. Further bleeding was observed in 25.5%, with a
patients with a life-threatening GI bleed or a supratherapeutic 7.2% incidence of thrombotic events and 30-day mortality of
INR substantially exceeding the therapeutic range. Its use could 7.0% (14,15,17–21). One study estimated transfusion-related
also be considered in those for whom massive blood transfusion events (fluid overload) of 4.9% within 7 days of PCC use (14).
is undesirable because of its effect on coagulopathy or dilution All studies demonstrated consistently rapid INR reduction of
of blood components when PCC is unavailable (see below). a large magnitude. Given the pharmacodynamics of warfarin
treatment, it was implausible that this dramatic INR change
2. For patients on warfarin who are hospitalized or under could have occurred because of bias, confounding, or chance.
observation with acute GI bleeding, we could not reach Studies were downrated for serious or very serious risk of
a recommendation for or against prothrombin complex bias (no comparator cohorts), indirectness of the outcome
concentrate administration. (“hemostatic efficacy” or active bleeding visualized at the
time of endoscopy), and the concomitant use of vitamin K.
Summary of evidence. The panel made an a priori decision to The small number of events contributed to serious impreci-
consider 3-factor PCC and 4-factor PCC equivalent for the sion. Only a small proportion of the patients had GI bleeds,
intervention (PCC for reversal of warfarin and other VKAs). No although the type of bleed would not have influenced the ef-
eligible studies were identified exclusively in patients with GI fect of PCC on the INR. Finally, the speed of INR correction
bleeding. A backward (snowballing) citation search of previous is a surrogate outcome, not a clinical outcome.
guidelines was used to identify supporting evidence, including
noncomparative cohort data derived from the PCC arms of 2 Conclusions. There is insufficient evidence to judge the
RCTs that compared PCC vs FFP (14,15). We also considered a balance between desirable and undesirable effects with
cohort study of GI patients, which provided clinical outcomes PCC administration; thus, the panel was unable to issue a
Table 2. Guideline statements, the strength of recommendation, and certainty of the evidence for the management of antithrombotic agents in the
elective endoscopy setting
Management of antithrombotic agents in the elective endoscopy setting
Anticoagulant interruption vs continuation

11. For patients on warfarin undergoing elective/planned endoscopic GI procedures, we suggest warfarin be continued, as opposed to temporarily
interrupted (1–7 d) (conditional recommendation, very low certainty of evidence).
12. For patients on warfarin, who hold warfarin in the periprocedural period for elective/planned endoscopic GI procedures, we suggest against
bridging anticoagulation (conditional recommendation, low certainty of evidence).
13. For patients on DOACs who are undergoing elective/planned endoscopic GI procedures, we suggest temporarily interrupting DOACs rather
than continuing DOACs (conditional recommendation, very low certainty of evidence).

Antiplatelet interruption vs continuation
14a. For patients on dual antiplatelet therapy for secondary prevention who are undergoing elective endoscopic GI procedures, we suggest tem-
porary interruption of the P2Y12 receptor inhibitor while continuing ASA (conditional recommendation, very low certainty of evidence).
14b. For patients on single antiplatelet therapy with a P2Y12 receptor inhibitor who are undergoing elective endoscopic GI procedures, we could
not reach a recommendation for or against temporary interruption of the P2Y12 receptor inhibitor.
15. For patients on ASA 81–325 mg/d (i.e., cardiac ASA monotherapy) for secondary prevention, we suggest against interruption of ASA (condi-
tional recommendation, very low certainty of evidence).
Timing of anticoagulant resumption after endoscopy
16. In patients who are undergoing elective endoscopic GI procedures whose warfarin was interrupted,we could not reach a recommendation for
or against resuming warfarin the same day vs 1–7 d after the procedure.
17. In patients who are undergoing elective endoscopic GI procedureswhose DOAC was interrupted, we could not reach a recommendation for
or against resuming the DOAC on the same day of the procedure vs 1–7 d after the procedure.
Timing of P2Y12 inhibitor resumption after endoscopy
18. In patients who are undergoing elective endoscopic GI procedures whose P2Y12 inhibitor was interrupted, we could not reach a recommen-
dation for or against resuming P2Y12 inhibitor on the same day of the procedure vs 1–7 d after the procedure.
ASA, acetylsalicylic acid; DOAC, direct oral anticoagulant; GI, gastrointestinal.
recommendation. The guideline panel implicitly considered FFP or PCC found that the absolute risk of further bleeding
evidence from the comparison of PCC with FFP for warfarin was numerically lower in the PCC arm with zero of 20
reversal that did reveal a favorable profile for PCC use and patients diagnosed with bleeding compared with 7 of 20
benefit in studies using the surrogate endpoint of INR correction. patients (35%) in the FFP arm but without statistical sig-
PCC is not necessary for most patients on warfarin with a GI nificance (RR 0.07, 95% CI: 0–1.09) (17). An additional 3
bleed. PCC administration could be considered in patients with patients in the FFP arm developed recurrent bleeding, but it
a life-threatening GI bleed, those with a supratherapeutic INR was unclear whether these 3 patients were independent of
substantially exceeding the therapeutic range, or in patients in the 7 patients already attributed. Regardless, their inclusion
whom massive blood transfusion is undesirable because of its would not change the direction of the effect nor certainty of
effect on coagulopathy or dilution of blood components. evidence.
The risk of thromboembolic events in patients on warfarin
3. For patients on warfarin who are hospitalized or under randomized to receive either FFP or PCC was evaluated in 2
observation with acute GI bleeding, we suggest pro- studies. In 1 study, the bleeding site was intracranial, whereas
thrombin complex concentrate administration com- in the second study, the bleeding site varied and included GI
pared with FFP administration (conditional recommen- bleeding in some patients (14,15). Combining results from both
dation, very low certainty of evidence). studies, although not significant, the absolute risk of thrombo-
embolic events was numerically higher in the PCC arm (RR =
Summary of evidence. We identified 2 randomized trials 1.60, 95% CI: 0.70–3.62), whereas the 30-day mortality (RR
(14,15) and 1 cohort study (17) comparing PCC with FFP in = 0.64, 95% CI: 0.17–2.49) and transfusion-related adverse
patients on warfarin with bleeding. The 2 studies that included events (1 transfusion-related anaphylaxis in the FFP group)
patients with GI bleeding reported inconsistent results (14,17). (14) were numerically lower in the PCC arm. Both studies
The RCT by Sarode et al. (14) did not find a difference in demonstrated a more rapid INR reduction in patients receiving
further bleeding among patients with acute GI bleeding with PCC than FFP (RR = 6.99, 95% CI: 3.61–13.53). The hetero-
PCC (25.4%) compared with FFP (24.1%) (relative risk [RR] geneous study populations, variability in outcome definition
1.05, 95% CI: 0.55–2.00). However, the study's definition of and timing of assessment, and the wide confidence intervals for
successful hemostasis allowed for up to 2 additional units of clinical outcomes led to a very low certainty of evidence.
blood products after receiving FFP or PCC. It did not report
whether this cointervention differed between the 2 groups. Conclusions. The effect of PCC compared with FFP on further
Furthermore, a higher proportion of patients in the FFP arm GI bleeding in patients on warfarin is unknown; however, the
received vitamin K, including intravenously. more rapid and reliable correction of the INR provides for a
A prospective cohort study of patients with acute upper biological rationale supporting the efficacy of PCCs. Although
GI bleeding who received intravenous vitamin K and either there was a very low certainty of evidence, the panel determined
that the anticipated desirable effects of PCC compared with FFP Summary of evidence. The available evidence addressing this
were greater than the undesirable effects in patients with acute recommendation included 1 cohort study that compared
GI bleeding. The panel concluded that although most patients idarucizumab with no treatment and 2 additional cohort studies
with acute GI bleeding on warfarin would not require PCC without a comparator. Singh et al. (24) performed a retrospective
administration, PCC use could be considered in patients with cohort study in the United States that included patients hospitalized
a life-threatening GI bleed, in those with a supratherapeutic for dabigatran-associated major nontraumatic GI bleeding or
INR substantially exceeding the therapeutic range, or those in intracranial bleeding. Among those with GI bleeding (159 who
whom massive blood transfusion is undesirable because of its received idarucizumab vs 1124 who did not), nonsignificant
effect on coagulopathy or dilution of blood components. differences in mortality (OR: 1.39, 95% CI: 0.51–3.45) and venous
thromboembolism (OR: 0.35, 95% CI: 0.08–1.58) were observed.
4. For patients on warfarin who are hospitalized or under In the Reversal Effects of Idarucizumab on Active Dabigatran
study (25,26), Pollack et al. examined patients on dabigatran with

observation with acute GI bleeding, we suggest against
the use of vitamin K (conditional recommendation, multiple causes of acute bleeding, including GI (45.5%) and other
very low certainty of evidence). patients on dabigatran about to undergo an urgent surgery or
procedure. The reversal of dabigatran anticoagulant effect (before
Summary of evidence. In patients receiving a VKA such as and up to 24 hours after the administration of idarucizumab) was
warfarin, low-dose oral vitamin K 1–2 mg can be used when assessed by dabigatran-specific coagulation function tests (dilute
there is an elevated INR (typically an INR ≥10) to restore thrombin time or ecarin clotting time), and the reduction in the
therapeutic-level anticoagulation (i.e., INR 2.0–3.0) (22). In the concentration of unbound dabigatran, both indirect measures of
setting of clinically significant GI bleeding requiring therapeutic the outcome of interest. In a subgroup analysis of 137 patients with
intervention, vitamin K 2–5 mg (oral or intravenous) reverses GI bleeding (27), Van der Wall et al. reported 30-day mortality and
anticoagulant effect (to INR ≤1.3) in 24–48 hours. Vitamin K thrombotic event rates of 11.1% and 3.6%, respectively, among
use does not achieve rapid hemostasis in patients with acute patients receiving idarucizumab.
bleeding (22). Consequently, the clinical value of vitamin K is
limited in most patients with acute GI bleeding, especially if Conclusions. Given the limited evidence of benefit and the high
the bleed is self-limiting, treatable through direct endoscopic cost of idarucizumab, the panel felt it could not recommend
hemostatic intervention, or if the INR is mildly elevated (e.g., routine use of idarucizumab for patients with GI bleeding
INR 1.5–2.5). Vitamin K can be administered in patients with who have taken dabigatran. However, selective use may be
a supratherapeutic INR if the intent is to reverse the effect appropriate in patients with a life-threatening GI bleed who
of a VKA over an extended period (i.e., 2–4 weeks) or if the have taken dabigatran within the past 24 hours.
objective is to stop the VKA altogether. This decision should be
undertaken in consultation with hematologists, cardiologists, or Reversal of rivaroxaban or apixaban with
other clinicians involved in patients' anticoagulant management. andexanet alfa
No prospective studies have assessed whether giving vitamin
K in VKA-treated patients with acute bleeding affects clinically 6. For patients on rivaroxaban or apixaban who are hos-
meaningful outcomes. In a meta-analysis involving nonbleeding pitalized or under observation with acute GI bleed-
patients on a VKA with a supratherapeutic INR, administra- ing, we suggest against andexanet alfa administration
tion of vitamin K was associated with small, nonsignificant (conditional recommendation, very low certainty of
increases in mortality (RR = 1.24; 95% CI: 0.62–2.47) and evidence).
thrombotic events (RR = 1.29; 95% CI: 0.35–4.78) (23). One
retrospective case-control study involving patients with VKA- Summary of evidence. Andexanet alfa, or “coagulation
associated bleeding (from both GI and non-GI sites) found factor Xa (recombinant) inactivated-zhzo,” is a modified
vitamin K administration was associated with a significant de- recombinant human factor Xa decoy protein that binds and
crease in mortality (adjusted OR = 0.47; 95% CI: 0.24–0.92). sequesters apixaban and rivaroxaban. It also binds and inhibits
However, this study had significant methodological limitations, tissue factor pathway inhibitor and has an elimination half-
including cointerventions that confounded an association be- life of 5 hours (28). In clinical trials, andexanet alfa decreased
tween vitamin K administration and clinical outcomes (13). apixaban activity by 94% and rivaroxaban activity by 92%.
It restored thrombin generation in 100% of patients within
Conclusions. Overall, there is no clinical evidence that vitamin 2–5 minutes.
K administration in VKA-treated patients with acute GI A prospective, single-group cohort of 352 patients with
bleeding prevents further bleeding or improves mortality or major bleeding within 18 hours of factor Xa inhibitor admin-
other clinically meaningful outcomes. Moreover, evidence is istration (rivaroxaban n = 128, apixaban n = 194, enoxaparin
very weak that giving or not giving vitamin K will affect the n = 20, or edoxaban n = 10) examined outcomes after giving
risk of thromboembolism, such as stroke or venous thrombosis, andexanet alfa. A subgroup (90 patients) suffered an episode
presumed to be because of normalization of the INR. of GI bleeding with all contributing to the ‟safety group,”
whereas 62 contributed to the ‟efficacy group” (those with a
Direct thrombin inhibitor reversal baseline anti-Xa activity of at least 75 ng/mL and confirmed
major bleeding) (29). Overall, the surrogate marker of me-
5. For patients on dabigatran who are hospitalized or dian anti-F Xa activity decreased from 149.7 to 11.1 ng/mL
under observation with acute GI bleeding, we suggest (92% reduction; 95% CI: 91–93) in the apixaban group and
against the administration of idarucizumab (condi- from 211.8 to 14.2 ng/mL (92% reduction; 95% CI: 88–94)
tional recommendation, very low certainty of evidence). in the rivaroxaban group.
Among the 62 patients with GI bleeding, excellent or selective use may be clinically justifiable in some patients
good hemostatic efficacy was noted 12 hours after the who have taken DOACs within the past 24 hours with a life-
andexanet alfa infusion in 85% (95% CI: 76–94), although threatening GI bleed.
the clinical applicability of the chosen criteria may not re-
flect contemporary clinical standards in GI bleeding. In add- Reversal of antiplatelet with platelet transfusion
ition, methodological limitations included the absence of an
intention-to-treat analysis, possible confounding covariates,
8. For patients on antiplatelet agents who are hospitalized
and insufficient reporting of resuscitative, endoscopic, and
or under observation with acute GI bleeding, we sug-
pharmacological management. Surprisingly, there was no
gest against platelet transfusions (conditional recom-
significant relationship between hemostatic efficacy and a re-
mendation, very low certainty of evidence).
duction in anti-FXa activity during andexanet alfa treatment.
Adverse events were reported only at the whole group level

Summary of evidence. ASA and the thienopyridine P2Y12
and included thrombotic events and mortality within 30 days receptor inhibitors clopidogrel and prasugrel irreversibly
in 9.7% of 352 patients and 13.9%, respectively. Infusion- block platelet function for the 7–10-day life span of platelets,
related events at 7 days were noted in 2 patients but not in the whereas ticagrelor is a reversible nonthienopyridine P2Y12
90 patients with GI bleeding. receptor inhibitor (a cyclopentyltriazolopyrimidine) that
impairs platelet function for 3–5 days. For this guideline, we
Conclusions. The only published study presents a serious risk
refer most frequently to thienopyridine antiplatelet agents
of bias because it lacks a control group. Indirectness of the
in discussing P2Y12 receptor inhibitors because the evidence
outcomes is also a significant concern because data on patients
reviewed examined clopidogrel or prasugrel. However,
with GI bleeding are limited, with missing information
the mechanism of action of the nonthienopyridine P2Y12
concerning specific management. Additional methodological
receptor inhibitor, ticagrelor, is similar, permitting reasonable
limitations include very serious imprecision as event rates are
extrapolation of results. Previous guidelines have suggested
low and the use of surrogate laboratory rather than clinical
platelet administration as a therapeutic option in patients on
outcomes. Notably, the cost of using the drug is high (up to
antiplatelet agents with severe GI bleeding (36,37). However,
$49,500 at high-dose regimen, with the low-dose regimen
the possibility of thrombotic events with an infusion of
costing half as much) (30). Accordingly, the panel could not
functional platelets in patients taking antiplatelet drugs, who
recommend the routine use of andexanet alfa in patients with
are at higher cardiovascular risk, along with potential risks
GI bleeding. This intervention could be considered in the
related to the transfusion of blood products, also needs to be
setting of life-threatening GI bleeding in hospitalized patients
considered.
who have taken apixaban or rivaroxaban within the past 24
hours. A single fully published study directly relevant to this
PICO was identified: a cohort study in patients without
thrombocytopenia taking antiplatelet agents and admitted
Reversal of DOACs with PCC with GI bleeding. This study compared 204 patients who
received platelet transfusion with a matched control group
7. For patients on direct oral anticoagulants who are hos- of 204 patients who did not. Adjusted analyses revealed a
pitalized or under observation with acute GI bleeding, significant increase in mortality (OR = 5.57, 95% CI: 1.52–
we suggest against prothrombin complex concentrate 27.1) and small, nonsignificant increases with platelet trans-
administration (conditional recommendation, very low fusion vs no transfusion in further bleeding (OR = 1.47, 95%
certainty of evidence). CI: 0.73–3.05) and thrombotic events (OR = 1.35, 95% CI:
0.74–2.49) (36).
Summary of evidence. The literature search identified only 2 Studies of platelet transfusion in patients for indications
cohort studies with comparator arms (no PCC) (31,32), both other than GI bleeding provide additional indirect evidence.
of which have limitations. Schulman et al. (31) examined An RCT of 190 patients with intracerebral hemorrhage re-
the reversal of dabigatran-associated major bleeding with ported an increase in the primary endpoint of death or de-
activated prothrombin concentrate in a small prospective pendence (because of significant neurological deficit) with
cohort study (N = 14). Among the 5 patients with GI bleeding platelet transfusion vs standard care (adjusted OR = 2.05,
compared with matched patients (N = 28) from 5 Phase III 95% CI: 1.18–3.56), as well as a small, nonsignificant in-
trials, the “effectiveness” rating was assessed at 24 hours by crease in mortality (RR = 1.38, 95% CI: 0.78–2.44) and a
the treating physicians for GI bleeding. The effectiveness was large, nonsignificant increase in thrombotic events (RR = 3.84,
considered good in 4 patients and moderate in 1, which was 0.44–33.68) with platelet transfusion (38). A retrospective
not statistically different from the comparator group. Smythe analysis of data from double-blind placebo-controlled RCTs
et al. (32) reported that among 28 patients with GI bleeding of patients undergoing coronary artery bypass graft sur-
on dabigatran, 2 received 4-factor PCC, and both (100%) gery reported higher mortality with platelet transfusion on
died within 30 days. However, the mortality results were not multivariable analysis with propensity scoring (OR = 4.76,
adjusted for confounding, and the comparator group's death 95% CI: 1.65–13.73) (39).
rate is unknown. Several systematic reviews (33–35) have
reported mainly on low-quality, single-arm cohort studies. Conclusions. Given a possible mortality increase in patients with
GI bleed and other medical conditions, and the lack of benefit
Conclusions. Given the uncertainty of the available evidence, in decreasing further bleeding in patients with GI bleeding,
the panel felt they could not recommend routine use of PCC for the panel suggests against platelet transfusion in patients with
patients with GI bleeding who have taken DOACs. However, antiplatelet-related GI bleeding who are not thrombocytopenic.
Holding ASA vs continuing ASA typical in current clinical practice) (46). Recurrent bleeding
rates at 30 days were not significantly greater in the ASA
9. For patients with GI bleeding on cardiac ASA for sec- group (10.3% vs 5.4%); difference 4.9% (95% CI: −3.6 to
ondary cardiovascular prevention, we suggest against 13.4), whereas 8-week mortality attributable to cardiovas-
holding the ASA (conditional recommendation, very cular, cerebrovascular, or GI complications was significantly
low certainty of evidence). greater in the placebo group (1.3% vs 10.3%; difference 9%
10.
For patients with GI bleeding on ASA for second- [95% CI: 1.7–16.3]). Thrombotic events at 30 days did not
ary cardiovascular prevention whose ASA was inter- differ between groups (3/78 vs 9/78 favoring early ASA re-
rupted, we suggest the ASA be resumed on the day sumption, RR = 0.33 [95% CI: 0.09–1.19]) with 6 nonfatal,
hemostasis is endoscopically confirmed (conditional recurrent acute ischemic events reported (2 in the ASA and
recommendation, very low certainty of evidence). 4 in the placebo group). However, the 2-month interruption
of ASA in the placebo group imparts serious indirectness of

observed outcomes. There also exists very serious imprecision
Summary of evidence. Current recommendations suggest that because of very low event rates.
patients with upper GI bleed undergo endoscopy within 24 Two additional cohort studies that compared patients who
hours, and in those with a lower GI bleed, diagnostic testing continued ASA after GIB to others who discontinued ASA
be performed within ∼24–36 hours (40–42). In addition, and did not resume ASA could not be included. In 1, the life-
hemostasis generally occurs before endoscopy or, in the table analysis curves did not permit accurate extraction of
minority with active bleeding identified endoscopically, at the results for the first 1–7 days (the relevant timeframe for this
time of endoscopy after hemostatic therapy is applied. ASA's recommendation) (47). The second study was unclear when
pharmacodynamic effect occurs through irreversible inhibition patients along the x axis were on or off ASA (48).
of platelet cyclooxygenase-1, which mediates thromboxane We wish to stress that our recommendations do not apply
synthesis. After ASA ingestion, thromboxane synthesis to patients taking ASA for primary cardiovascular prevention.
normalizes by 7–10 days, although in vitro studies suggest 70% Recent RCTs suggest little if any benefit of primary preven-
of arachidonic acid–mediated platelet function may normalize tion for reduction of cardiovascular outcomes despite sig-
by 3 days after ASA ingestion (43,44). Therefore, interruption nificant increases in serious GI bleeding (49–51), and current
of ASA in patients presenting with GI bleeding will have little guidelines suggest ASA for primary prevention be considered
impact on the initial clinical course because of the persistent only in a very limited population and should not be used in
antiplatelet effect of ASA in the first day or 2 after the patient's those with increased risk of bleeding (52,53).
presentation. Furthermore, the suggestion to resume ASA
immediately after hemostasis means that ASA resumption Conclusions. The panel weighed the important and well-
will occur in most cases well before the antiplatelet effect has documented cardiovascular benefit of secondary preventive
substantially waned. Thus, the initial interruption of ASA after ASA therapy and the potential risk of further GI bleeding with
presentation would not be expected to have much impact on continued ASA therapy. The trend to reduced mortality in an
either bleeding or cardiovascular clinical outcomes if ASA is observational study of patients with myocardial infarction with
restarted once endoscopic hemostasis is established. continued aspirin (45) coupled to the significant reduction in
Only 1 study was identified relevant to the PICO regarding mortality among patients with high-risk ulcer bleeding who
interruption vs continuation of ASA when patients present had aspirin resumed immediately after endoscopic hemostasis
with GI bleeding (GIB); this was a retrospective study in (46) were important considerations in the panel's decision to
patients hospitalized with acute myocardial infarction who recommend continuation rather than an interruption of aspirin
subsequently developed ulcer bleeding during hospitaliza- therapy. If ASA is discontinued at clinical presentation, we
tion (45). This comparison of 64 patients interrupting ASA recommend rapid resumption within 24 hours of successful
vs 38 continuing ASA reported similar 30-day rates for both endoscopic hemostasis.
further bleeding and for mortality of 16% (10/64) vs 11% Increased further bleeding with continued ASA at presen-
(4/38) (RR = 1.48, 0.50–4.41), respectively. The results for tation was not shown in the observational study, but results
further bleeding are opposite than expected for interruption of the RCT did raise the possibility of increased rebleeding
of ASA (i.e., slightly more bleeding reported with interrupted with early resumption of ASA in patients with high-risk ulcer
ASA), although confidence intervals are wide and consistent bleeding. The panel also weighed the preferences of providers
with benefit or harm. No statistical adjustments were made and patients for a cardiovascular event vs a GIB event, as
for potential confounders, limiting the utility of the results. discussed in Supplementary Digital Content (see Appendix
Furthermore, since outcomes were at 30 days, the results may 2, http://links.lww.com/AJG/C417) in formulating their
be viewed as primarily relevant to the timing of ASA resump- recommendation.
tion after bleeding stops.
The most pertinent study relevant to the PICO regarding Management of antithrombotic agents in the
resumption of ASA after hemostasis is an RCT in patients elective endoscopy setting
taking ASA for secondary cardiovascular protection with The 9 remaining statements inform antithrombotic therapy
high-risk ulcer bleeding requiring endoscopic therapy. In this management in patients undergoing scheduled, elective
RCT, 156 patients with peptic ulcer bleeding and high-risk endoscopic procedures. These recommendations exclude
endoscopic stigmata treated with successful endoscopic patients at high risk of thromboembolic events in whom
therapy and proton pump inhibitor were randomized to con- elective procedures should be deferred. Such high-risk
tinued low-dose ASA for secondary prevention vs placebo for patients include those within 3 months of acute venous
the 8 weeks of the study (a much longer interruption than thromboembolism (comprising lower-limb deep vein
thrombosis or pulmonary embolism), stroke, or transient preferences (63,64). Indeed, patients placed more weight
ischemic attack (Table 4); and patients within 3 months of (more disutility) on stroke prevention than GI bleeding un-
acute coronary syndrome (ACS) event, within 6 months of less they had previously experienced a GI bleed (64). Among
a drug-eluting stent or 1 month of a bare-metal coronary the latter, 87% placed the highest utility on rebleeding risk
stent placement without ACS history (54); or after ACS followed by thrombosis risk (64). The panel members con-
event within 12 months of a drug-eluting stent placement or cluded that for most PICOs, there is possibly significant un-
2 months of bare-metal stent placement (54). Recent data certainty about or variability in how much people value the
suggest that dual antiplatelet therapy (DAPT) with ASA critical outcomes.
and P2Y12 receptor inhibitor can be converted to platelet
P2Y12 receptor inhibitor monotherapy among patients at 3 Anticoagulant interruption vs continuation
months or less in patients with a drug-eluting stent placed
after ACS event (55,56). 11. For patients on warfarin undergoing elective/planned

A review of published guidelines highlights the lack of con- endoscopic GI procedures, we suggest warfarin be
sensus regarding high vs low baseline risk of endoscopic pro- continued, as opposed to temporarily interrupted
cedures (Table 3) (37,57–59). Procedural bleeding risk and (1–7 days) (conditional recommendation, very low
patient-specific thromboembolic risk were empirically framed certainty of evidence).
using the risk stratification endorsed by the International
Society on Thrombosis and Haemostasis Guidance Statement Summary of evidence. The published data examining
(60), the BRIDGE Trial (61), previously published guidelines, uninterrupted warfarin before endoscopic procedures and
and expert opinion (Table 4) (37,57–59,61). A structured and associated GI bleeding risk is heterogeneous and methodologically
exhaustive GRADE assessment of procedural bleeding risk is flawed. When formulating our recommendation, we considered
beyond the scope of this clinical practice guideline. 3 cohort studies with a control group (temporary interruption of
Also pertinent to this section are considerations of patient warfarin) (65–67) and 2 cohort studies without a control group
preference. The targeted review performed for this guideline (68,69). These 5 studies provided very low certainty of evidence
initiative (see Appendix 2, Supplementary Digital Content, because of imprecision of the results, lack of adjustment for
http://links.lww.com/AJG/C417) demonstrated substantial confounders, lack of standardized procedure technique (e.g.,
variability in the threshold number of bleeds observed for biopsy, cold snare polypectomy, hot snare polypectomy, and
oral anticoagulation therapy to be considered acceptable routine hemostatic clipping), comparator groups that differ in
both within individuals and between different studies (62). prognostic factors for bleeding, differences in population, lack
Furthermore, country-specific differences exist in patients' of blinding of the endoscopist, and incomplete follow-up.
perceptions of atrial fibrillation, concerns about stroke, and The desirable anticipated effect with continued warfarin
preference for involvement in oral anticoagulation therapy (compared with interrupted warfarin) is reduced thrombo-
treatment decisions, with recent experience of stroke and GI embolic events. A single small cohort study without ad-
bleeding both significantly influencing patient values and justment for confounding factors reported a nonsignificant
Table 3. Empiric endoscopic procedural bleeding risk stratification
High bleeding risk procedures (30-d risk of major bleed >2%) Low/moderate bleeding risk procedures (30-d risk of major bleed ≤2%)
Polypectomy (≥1 cm) EGD with/without biopsy

PEG/PEJ placement Colonoscopy with/without biopsy
ERCP with biliary or pancreatic sphincterotomy Flexible sigmoidoscopy with/without biopsy
EMR/ESD ERCP with stent (biliary or pancreatic) placement or papillary balloon
dilation without sphincterotomy
EUS-FNA EUS without FNA
Endoscopic hemostasis (excluding APC) Push enteroscopy and diagnostic balloon-assisted enteroscopy
Radiofrequency ablation Enteral stent deployment
POEM Argon plasma coagulation
Treatment of varices (including variceal band ligation) Balloon dilation of luminal stenoses
Therapeutic balloon-assisted enteroscopy Polypectomy (<1 cm)
Tumor ablation ERCP without biliary or pancreatic sphincterotomy
Cystogastrostomy Marking (including clipping, electrocoagulation, and tattooing)
Ampullary resection Video capsule endoscopy
Pneumatic or bougie dilation
Laser ablation and coagulation
The sources used for the empiric classification of procedures included the International Society on Thrombosis andHaemostasis Guidance Statement, the
BRIDGE trial, previously published guidelines, and expert opinion by the authors.
APC, argon plasma coagulation; EGD, esophagogastroduodenoscopy; EMR, endoscopic mucosal resection; ERCP, endoscopic retrograde
cholangiopancreatography; ESD, endoscopic submucosal dissection; EUS, endoscopic ultrasound; FNA, fine-needle aspirate; PEG, percutaneous endoscopic
gastrostomy; PEJ, percutaneous endoscopic jejunostomy; POEM, peroral endoscopic myotomy.
Table 4. Empiric periprocedural thromboembolic risk stratification for patients receiving anticoagulant therapy
Risk Indication for anticoagulation

stratum
Mechanical heart valve Atrial fibrillation Venous thromboembolism
Higha • Any mitral valve prosthesis • CHADS2 score: 5 or 6 • Recent (within 3 mo) VTE
• Any caged-ball or tilting disc aortic • CHA2DS2VaSc score: ≥7 • Severe thrombophilia (e.g., defi-
valve prosthesis • Recent (within 3 mo) stroke or transient ciency of protein C, protein S or
• Recent (within 3 mo) stroke or ischemic attack antithrombin, antiphospholipid anti-
transient ischemic attack • Rheumatic valvular heart disease bodies, and multiple abnormalities)
Moderate • Bileaflet aortic valve prosthesis and • CHADS2 score: 2–4 (no previous stroke • VTE within the past 3–12 mo
≥1 of the following: atrial fibrilla- or transient ischemic attack) • Nonsevere thrombophilia (e.g.,

tion, previous stroke or transient • CHA2DS2VaSc score: 5 or 6 heterozygous factor V Leiden or pro-
ischemic attack, hypertension, thrombin gene mutation)
diabetes, congestive heart failure, • Recurrent VTE
and age ≥75 yr • Active cancer (treated within 6 mo or
palliative)
Low • Bileaflet aortic valve prosthesis • CHADS2 score: 0 or 1 • VTE more than 12 mo ago and no
without atrial fibrillation and no • CHA2DS2VaSc score: 1–4 other risk factors
other risk factors for stroke
The sources used for the empiric classification of procedures included the International Society on Thrombosis andHaemostasis Guidance Statement, the
BRIDGE trial, previously published guidelines, and expert opinion by the authors. VTE, venous thromboembolism.
a
High-risk patients may also include patients with a previous stroke or transient ischemic attack occurring >3 mo ago and a CHADS2 score <5, patients
with previous thromboembolism during temporary interruption of VKAs, or those patients undergoing certain types of surgery (e.g., cardiac valve
replacement, carotid endarterectomy, and major vascular surgery).
reduction with uninterrupted vs interrupted warfarin with advanced endoscopic procedure (Table 3), which may incur
0/43 vs 1/19 thromboembolic events, respectively (RR 0.15; a higher risk of procedural bleeding, in which case 5 days of
95% CI: 0.006–3.56) (67). temporary interruption without bridging heparin would be
Our ability to estimate the direction and magnitude of the appropriate, as discussed in PICO 12.
effect of uninterrupted warfarin (compared with interrupted
warfarin) on GI bleeding and mortality using data from the 3 12. For patients on warfarin, who hold warfarin in the
cohort studies with controls (65–67) is limited by small sample periprocedural period for elective/planned endo-
size and few events, resulting in extremely wide confidence scopic GI procedures, we suggest against bridging
intervals compatible with considerable benefit and consider- anticoagulation (conditional recommendation, low
able harm. When the continuous warfarin arms from all 5 certainty of evidence).
studies are pooled (65–69), we observe 0/239 bleeding events
(95% CI: 0%–12.5%), suggesting a risk of postprocedural Summary of evidence. In patients receiving warfarin who
bleeding with continuous warfarin as low as 0% and as high require its temporary interruption, heparin bridging, typically
as 12.5%. with subcutaneous, full-dose low-molecular-weight heparin
(LMWH), is sometimes used for 3 days before and 3–5 days
Conclusions. It is impossible to confidently estimate the GI after the surgery or procedure. The premise for heparin bridging
procedural bleeding risk associated with uninterrupted warfarin is that by minimizing the time patients are not therapeutically
therapy (vs warfarin interruption), given the limitations of the anticoagulated periprocedurally during warfarin interruption
published literature, heterogeneity of patient populations and and resumption, the risk of stroke and other thromboembolic
procedure type, and imprecision of the results. The absence events will be reduced. However, heparin bridging may
of studies in advanced endoscopic procedures (with higher not affect the pathophysiologic pathway that mediates
baseline bleeding risk), and differences in clinical consequences periprocedural stroke and thromboembolism and may place
of luminal and extraluminal bleeding associated with such patients at increased risk of procedure-site bleeding, especially
endoscopic GI procedures, limits our ability to comment on if heparin bridging is administered in too close proximity to the
the safety of proceeding without interrupting warfarin. The time of the procedure (70,71).
evolving role of mechanical hemostasis may render some Two randomized trials assessed heparin bridging among
advanced procedures safer with continued warfarin in the warfarin-treated patients who required anticoagulant inter-
future; however, current evidence supporting this strategy is ruption for elective surgery/procedure, including GI proced-
scant. ures. One randomized, double-blind, placebo-controlled trial
The planned procedure type (Table 3) and its associated (BRIDGE) assessed the need for heparin bridging in patients
risk of postprocedural bleeding, and the baseline risk of with atrial fibrillation who required temporary warfarin
thromboembolism will influence the recommendation, as interruption for an elective surgical procedure, including
will resource requirements associated with discontinuation 758 GI procedures (98.7% were minor or low bleeding risk
and reinitiation of anticoagulation (e.g., laboratory tests and procedures) (72). Excluded from the BRIDGE trial were pa-
clinic visits). For patients on warfarin who are undergoing tients with a very high thrombotic risk (i.e., a mechanical
elective and planned outpatient endoscopic GI procedures, we heart valve; stroke, systemic embolism, or transient ischemic
suggest warfarin be continued unless they are undergoing an attack within the past 12 weeks) or significant risk factors
for major bleeding (i.e., history of a bleeding event within study (77), the incidence of 30-day thrombotic events and
the past 6 weeks; creatinine clearance <30 mL/min; and mortality was 0.7% and 0.5%, respectively, after DOAC
thrombocytopenia < 100,000 per microliter). Patients with temporary interruption (77).
planned cardiac, intracranial, or intraspinal surgeries were The absolute risk of increased delayed bleeding with con-
also excluded (72). tinuous DOAC anticoagulation could not be reliably calcu-
There were 1,813 patients enrolled in the BRIDGE lated nor the results reliably pooled, given the zero event rates
trial, of whom 918 were allocated to receive bridging with in one or both arms of comparative studies (65,67,69,76), the
therapeutic-dose LMWH before and after the surgical pro- unclear denominator for patient numbers (76), and the ab-
cedure, and 895 to matching placebo, with a 30-day follow-up sence of control arms (69,77). Furthermore, we noted a lack
period after procedure. Forgoing bridging anticoagulation of adjustment for known confounders (65,67,76), limited
was noninferior to perioperative bridging with LMWH for sample sizes, and low event rates. In addition, there was a
the prevention of arterial thromboembolism (0.4% vs 0.3%,

diversity of GI procedure types, endoscopic techniques, and
respectively, with a risk difference, of 0.1%; 95% CI: −0.6 protocols for DOAC interruption. These factors contribute to
to 0.8; P = 0.01 for noninferiority) and decreased the risk of the serious risk of bias, inconsistency, indirectness, and im-
major bleeding (1.3% vs 3.2%, respectively, P = 0.005) (72). precision in estimates, leading to the very low certainty of
Another randomized trial of bridging (Postoperative low evidence.
molecular weight heparin bridging treatment for patients at The most informative study was the prospective PAUSE
high risk of arterial thromboembolism [PERIOP-2]) was per- cohort study (77), which provided a standardized protocol
formed in 1,471 warfarin-treated patients who required an for DOAC interruption, complete follow-up, and valid out-
elective surgery or procedure in which all patients received come assessment. However, there was no comparator of
preprocedure LMWH bridging and were randomly allocated interest (i.e., uninterrupted anticoagulation). The panel was
to receive bridging, with either a therapeutic-dose or low-dose provided with the raw data for the subgroup of patients
LMWH regimen (determined by the procedure bleed risk) or undergoing GI procedures enrolled in this cohort study. The
no bridging after procedure (73). This study was unique in calculated GI bleeding incidence rate was 2.5% (95% CI:
that it included patients with mechanical heart valve (n = 1.4%–4.2%), with 35.7% of GI bleeding events considered
304) in addition to patients with atrial fibrillation (n = 1,167). major bleeding episodes (Jim Douketis, Alan Barkun, written
PERIOP-2 was not included in the evidence profile because it communication, May 15, 2021). The 30-day thrombo-
had only been published in abstract form. A few weeks after embolic (0.7%; 95% CI: 0.3%–1.8%) and mortality (0.5%;
the final panel voting meeting, the PERIOP-2 trial was pub- 95% CI: 0.2%–1.6%) incidence rates were also very low
lished as full text showing similar results to the BRIDGE trial and nonsignificant with temporary DOAC interruption. Of
(72). Two additional observational studies of lower methodo- the 556 endoscopic procedures performed, most were col-
logical quality involving only warfarin-treated patients who onoscopies, gastroscopies, and flexible sigmoidoscopies with
required an elective GI procedure further suggest that the and without biopsy or polypectomy. Before endoscopic pro-
use of periprocedural heparin bridging increases the risk of cedures, the duration of DOAC interruption was 2.0 ± 0.5
postprocedure bleeding (74,75). days (including the day before the procedure and the day
of the procedure in 91.7%). Only 8.1% of patients under-
Conclusions. Overall, evidence is lacking that routine going scheduled endoscopic procedures held their DOACs
periprocedural heparin bridging during VKA interruption for >2 days before the procedure. DOAC resumption after
provides a therapeutic benefit to reduce thromboembolism procedure was 1.9 ± 1.5 days providing endoscopic hemo-
and seems to increase patients' risk of postprocedural bleeding. stasis had been achieved, for a total time off DOACs of 3.9
Periprocedural bridging may be appropriate in the subset of ± 1.6 days in the periendoscopic period (Jim Douketis, Alan
patients with mechanical valves, atrial fibrillation with CHADS2 Barkun; personal communication).
score >5, patients with previous thromboembolism during
temporary interruption of VKAs, or those patients undergoing Conclusions. From the limited available data, for patients on
certain types of surgery (e.g., cardiac valve replacement, carotid DOACs undergoing elective, planned endoscopic GI procedures,
endarterectomy, and major vascular surgery). Consultation we suggest temporary interruption of the DOACs is preferred
with a cardiologist and hematologist is recommended in these over continued DOAC administration. The duration of
high-risk thromboembolic patients. temporary DOAC interruption before endoscopic procedures
associated with favorable outcomes is between 1 and 2 days,
13. For patients on direct oral anticoagulants (DOACs) excluding the day of the procedure, which permits the shortest
who are undergoing elective/planned endoscopic preprocedural duration of DOAC interruption while balancing
GI procedures, we suggest temporarily interrupting bleeding and thromboembolism risk.
DOACs rather than continuing DOACs (conditional As the window of temporary interruption evaluated in this
recommendation, very low certainty of evidence). clinical question was 1–5 days before endoscopy, the panel
discussed if withholding DOACs for 1–5 days could trigger
a prothrombotic state that might result in thrombosis with
Summary of evidence. No RCTs addressed this clinical any subsequent postendoscopic delays in DOAC resumption.
question. However, 3 cohort studies with control arms It was argued that the prothrombotic risks seem to be more
(65,67,76) and 2 cohort studies without control arms related to the periprocedural milieu (e.g., nature of the inter-
(69,77) were used to indirectly estimate the risk of GIB with vention such as vascular surgery vs nonvascular surgery and
continuous DOAC anticoagulation (3.6%; 8/224) vs that patient characteristics) than the brief interruption of DOACs
with temporary interruption (3.1%; 18/578). In the PAUSE (78). Furthermore, given the rapid action of onset and half-life
of DOACs, the thrombotic risk of interruption is anticipated Summary of evidence. We identified 1 randomized trial and 1
to be lower than with interruption of warfarin (61,79). cohort study evaluating patients on P2Y12 inhibitors undergoing
elective endoscopic procedures that compared interruption
Antiplatelet interruption vs continuation with the continuation of antithrombotic therapy (80,83). In
their study, Chan et al. randomized 216 patients on clopidogrel,
14A. For patients on dual antiplatelet therapy for second- with or without concomitant ASA, to continued medication
ary cardiovascular prevention who are undergoing or placebo. The method of polypectomy included cold snare,
elective endoscopic GI procedures, we suggest tem- hot snare, cold biopsy, and hot biopsy without a prophylactic
porary interruption of the P2Y12 inhibitor while con- clip or endoscopic loop placement. None of the 46 patients on
tinuing ASA (conditional recommendation, very low clopidogrel alone were diagnosed with GIB; however, the study
certainty of evidence). was underpowered to detect a difference in this subgroup (80).

A retrospective cohort study of 1,050 patients on
Summary of evidence. The panel considered 2 RCTs and antiplatelet therapy undergoing colonoscopy with hot
numerous observational studies that examined the temporary snare polypectomy and, in most cases, prophylactic clip
interruption of DAPT (stopping the P2Y12 inhibitor while placement included 37 patients receiving P2Y12 inhibitors
continuing ASA) in patients undergoing elective endoscopic GI (83). The authors reported that 3 of 18 patients (16.7%,
procedures. Chan et al. (80) conducted a double-blinded RCT 95% CI: 4.4–42.3) who interrupted P2Y12 inhibitor
to examine the bleeding and thrombosis rates among patients therapy developed bleeding compared with none of the 19
treated with either clopidogrel 75 mg or placebo for 7 days patients who continued therapy (RR 7.37, 95% CI: 0.41–
before the colonoscopy (N = 387; of which N = 216 had cold 133.37). This study was limited by potential confounding
snare polypectomy). Of the 387 patients enrolled, 78.5% were because patient factors may have determined whether
on continuous ASA. There were similar rates of immediate and P2Y12 inhibitor was interrupted and the behavior of the
delayed postpolypectomy bleeding and a modest trend toward endoscopist during the colonoscopy. Both studies reported
fewer cardiothromboembolic events with thienopyridine thromboembolic events. There was a single event in the
interruption: 1.3% (95% CI: 0.3%–5.0%) of patients receiving interrupted P2Y12 inhibitor arm of 38 patients (2.6%, 95%
placebo vs 2.7% (95% CI: 1.0%–7.0%) in those with continued CI: 0.4–17.3) compared with none of the 58 patients con-
clopidogrel; RR = 0.47 (95% CI: 0.09–2.55). By contrast, Won tinuing therapy (RR 4.54, 95% CI: 0.19–108) (80,83). One
et al. (81) reported no thrombotic events with similar bleeding study reported 30-day mortality, and there were no deaths
rates among 87 patients randomized to continue DAPT through in either group (83).
a scheduled cold snare polypectomy (0/45 with placebo and
1/42 [2.4%] with DAPT). Conclusions. Although interruption of a P2Y12 inhibitor
The numerous observational studies examining should decrease a patient's risk of bleeding, the available
periprocedural antiplatelet regimens have been summarized in evidence reported a nonsignificant increased bleeding risk in
a systematic review by Eisenberg et al. (82). They assessed the patients who stop a P2Y12 inhibitor for an elective endoscopic
time to late stent thrombosis (occurring between 30 days and procedure compared with those who continue the medication.
1 year after stent implantation) in patients with drug-eluting This result is biologically implausible and, coupled to the very
stents on DAPT after discontinuing thienopyridine alone or large confidence intervals, speaks to the very low certainty
discontinuing both thienopyridine and ASA. However, the of evidence. Ultimately, the panel was unable to make a
absolute risk of stent thrombosis within 10 days when a recommendation.
P2Y12 inhibitor is discontinued while continuing ASA cannot
be calculated because the denominator (patients at risk) is 15.
For patients on ASA 81–325 mg/d (monotherapy)
unknown. Nonetheless, among the 94 patients with stent for secondary cardiovascular prevention, we suggest
thrombosis after discontinuing a P2Y12 inhibitor but con- against interruption of ASA (conditional recommen-
tinuing ASA, only 6 cases (6%) occurred within 10 days, sug- dation, very low certainty of evidence).
gesting that late stent thrombosis is a greater problem than
immediate stent thrombosis. There were no events reported Summary of evidence. The risk of clinically significant bleeding
in the 3–4 days after coronary intervention in this study. We with diagnostic endoscopic procedures and standard biopsies
note that this PICO considered only data regarding events is so low that the panel agreed that ASA does not need to be
occurring in the first 30 days after the intervention. held for these procedures (Table 3). A prospective observational
study of the risk of clinical bleeding (>2-g/dL hemoglobin drop
Conclusions. Among patients on DAPT (P2Y12 inhibitor necessitating endoscopic hemostasis) after endoscopic biopsies
[clopidogrel, prasugrel, or ticagrelor and ASA 81–325 mg/d]) revealed bleeding events in 0 of 142 patients continuing ASA and
for secondary cardiovascular prevention, we suggest temporary 1 of 61 (1.6%) interrupting ASA (65). Using the ASA arm of
interruption of the P2Y12 inhibitor. This recommendation an RCT comparing clopidogrel with ASA in healthy volunteers
applies only to elective and not emergency procedures. undergoing duodenal and antral biopsies, none of the 280
biopsies on ASA led to bleeding events (84).
14B.
For patients on single antiplatelet therapy with The risk of bleeding with polypectomy is higher than biop-
P2Y12 inhibitor agents who are undergoing elective sies, especially for larger polyps and with the use of cautery
endoscopic GI procedures, we could not reach a rec- rather than a cold snare. A case-control study examined 81
ommendation for or against temporary interruption patients with postpolypectomy bleeding matched to 81 pa-
of the P2Y12 inhibitor. tients who had polypectomies without complication. In this
study, 87% of polypectomies were performed with cautery, most polypectomies. Nevertheless, when removing larger
and 3% of polyps were >10 mm in size (85). ASA use within and more complex polyps and other procedures with
3 days before polypectomy exhibited a small, nonsignificant the highest bleeding risk (e.g., ESD, biliary or pancreatic
trend to being more common in the bleeding group (40% sphincterotomy, ampullectomy, peroral endoscopic myotomy,
vs 33%; OR = 1.41, 95% CI: 0.68–3.04). Several factors and radiofrequency ablation), the panel felt that interruption
limit the utility of this study, including the indirectness of the of ASA could be considered. Such decisions require consid-
population studied and lack of adjustment for confounding eration of other factors such as cardiovascular risk and pa-
factors. For example, the number of polyps removed was tient preference regarding cardiovascular vs bleeding events.
∼2-fold higher in the postpolypectomy bleeding group, which Patients taking ASA as primary prevention should have ASA
might suggest ASA use was less commonly associated with stopped before higher-risk endoscopic procedures because the
postpolypectomy bleeding when assessed on a per-polyp bleeding risk outweighs the minimal cardiovascular benefit.
rather than per-patient basis. The cases and controls were

derived from different databases, and the data are not gen- Timing of anticoagulant resumption after
eralizable to current practice in which most polyps <10 mm endoscopy
are removed with a cold snare. Furthermore, it is not clear
that DAPT was excluded. An observational study with infor- 16. In patients who are undergoing elective endoscopic GI
mation on a control group of 297 patients undergoing 867 procedures whose warfarin was interrupted, we could
polypectomies (mean size of largest polyp 6.5 mm; 29% hot not reach a recommendation for or against resuming
snare, 4% cold snare, and 71% cold forceps) revealed delayed warfarin the same day vs 1–7 days after the procedure.
rebleeding in 0 (95% CI: 0%–3%) of the 119 patients on ASA
monotherapy (86). Summary of evidence. The appropriate timing of warfarin
Procedures with the highest bleeding risk include wide- resumption after an elective endoscopic procedure is not known.
field endoscopic mucosal resection, endoscopic submucosal No prospective trials exist comparing different strategies. Three
dissection (ESD), biliary or pancreatic sphincterotomy, and single-arm prospective cohort studies were identified that
ampullectomy (Table 3). A retrospective study in patients reported outcomes of interest, grouped by the timing of warfarin
undergoing gastric ESD revealed bleeding in 1 (1.9%) of 53 resumption. Douketis et al. (93) evaluated 650 consecutive
patients interrupting ASA 7 days before ESD vs 2 (16.7%) of patients who required interruption of warfarin for an invasive
12 patients continuing ASA (RR 0.11; 95% CI: 0.01–1.15) procedure, including 5 patients undergoing colonoscopic
(87). No RCTs directly relevant to this PICO were identi- polypectomy and 65 patients undergoing endoscopy with or
fied on our search. Recently, an RCT of 552 patients under- without biopsy. All patients resumed warfarin on the procedure
going gastric ESD was published in abstract form after the day and received standardized LMWH bridging therapy. The
final panel voting meeting showing similar results (88). As the endoscopic procedures were analyzed as part of a subgroup
study was available only in abstract form, it was not included of 542 patients undergoing non–high bleeding risk procedures
in the formal evidence report. (e.g., cholecystectomy, bowel resection, angiography, and joint
Given the extremely limited evidence from studies in GI replacement). There were 4 (0.74%; 95% CI: 0.20–1.47) cases
bleeding, especially on our critical outcome of thrombotic of major bleeding, none of which were GI bleeding.
events, we also assessed studies in non-GI procedures to
Dunn et al. (94) studied 260 patients, including 46 GI inva-
assess the impact of ASA interruption vs continuation on
sive procedures, who resumed warfarin the day of their pro-
thrombotic events. A meta-analysis of 4 RCTs in patients
cedure. All patients received a standardized LMWH bridging
undergoing noncardiac surgery revealed a nonsignificant in-
therapy. One of the 46 patients (2%; 95% CI: 0–13) under-
crease in thrombotic events (RR = 1.49; 95% CI: 0.56–3.96)
going colonoscopy was diagnosed with GI bleeding during
with ASA interruption (89–92). Meta-analysis with these 4
the follow-up period of 28 days. Paik et al. (95) reported
RCTs and a 5th RCT (99) revealed a nonsignificant decrease
on 96 patients undergoing endoscopic sphincterotomy who
in postprocedural bleeding when ASA is interrupted (RR =
interrupted their warfarin therapy before the procedure and,
0.81; 95% CI: 0.66–1.01); the panel did not believe that
we assume, resumed warfarin on the day of the procedure.
bleeding rates with surgery could be generalized to GI endo-
Patients received different regimens of bridging therapy with
scopic procedures.
heparin. The study may be further limited in generalizability
because of the high rate of biliary stent placement (75% of
Conclusions. The panel weighed the potential desirable cases) and precut sphincterotomy (23% of cases). Including
effects (reduction in thrombotic events) and undesirable the 6 patients excluded for postprocedural bleeding before
effects (increased bleeding) of continuing ASA, limited by the the resumption of heparin, 11 of 102 patients (11%; 95%:
availability of only scant, very low certainty evidence. The CI 6%–19%) had postprocedure bleeding within the 14 days
known important benefit of ASA for secondary cardiovascular after endoscopic retrograde cholangiopancreatography.
prevention and the possible reduction in thrombotic events seen Thromboembolic events were reported in all 3 studies but
in RCTs of nonendoscopic surgical procedures led the panel to assessed at different times. There were 2 thromboembolic
conditionally suggest the continuation of ASA for endoscopic events in the 542 patients (0.37%; 95% CI: 0.04%–1.32%)
procedures, in general. However, a blanket recommendation undergoing non–high bleeding risk procedures at a median
cannot be made for all procedures and patients, given that of 13.8 days, including 1 event in a patient undergoing en-
bleeding risk varies markedly among endoscopic procedures, doscopy (93). Dunn et al. (94). reported 5 thromboembolic
and cardiovascular risk also varies among patients. events in 260 patients (1.9%; 95% CI: 0.6%–4.4%) within
The panel felt comfortable that the bleeding risk was 28 days of their procedure, none occurring in the 46 who
very low in diagnostic endoscopic procedures, biopsies, and underwent GI procedures. Finally, among the 96 patients
undergoing endoscopic sphincterotomy, 3 thromboembolic the procedure. Fourteen patients developed GI bleeding after
events (3.4%; 95% CI: 0.7%–8.9%) were diagnosed within endoscopy (2.5%, 95% CI: 1.4–4.2; n = 554) during the 30
the 30 days after procedure (95). Two studies reported mor- days of follow-up after the resumption of DOACs, of which 5
tality, and there were no deaths in either study at 13.8 or 28 were considered major bleeding episodes (77).
days, respectively (93,94). Radaelli et al. (96) reported thromboembolic events in 1 of
The lack of a comparator group limited these studies, as 477 patients resuming DOACs on day 0–3 and 1 of 52 patients
did the diversity of the populations studied, the small propor- resuming DOACs after day 3 (RR 0.11; 95% CI: 0.01–1.57).
tion of patients undergoing endoscopic procedures, the use Douketis et al. (77) reported an overall rate of thrombo-
of bridging therapy, and the outcomes assessment occurring embolic events of 21 in the entire cohort of 3,007 (0.7%; 95%
at variable follow-up times. The 2016 American Society for CI: 0.45%–1.09%) and 5 in the subgroup of patients under-
Gastrointestinal Endoscopy clinical practice guidelines (37) going GI endoscopic procedures (0.7%; 95% CI: 0.3%–1.8%).
recommend resuming warfarin the day of an elective endo- Mortality ranged from 0 in both arms of the Radaelli study

scopic procedure while referencing 2 studies lacking valid to 0.3% (95% CI: 0.15%–0.59%) among all patients in the
comparator arms and unclear resumption timing; these PAUSE cohort, to 0.5% (95% CI: 0.2%–1.6%) in the post hoc
studies were excluded from the current evidence profile. analysis of the GI PAUSE data (77,96). Previously published
clinical practice guidelines have made informal and formal re-
commendations regarding DOAC resumption after elective
Conclusions. We could not find studies comparing same-day
endoscopic procedures; however, none were informed by studies
resumption of warfarin with resumption in 1–7 days after
that evaluated the timing of DOAC resumption (37,59,97).
the temporary interruption of warfarin before an elective
endoscopic procedure. Therefore, the panel was unable to make
a recommendation. In PICO 11, the panel suggested continuing Conclusions. We did not identify a study comparing the timing
warfarin in patients undergoing elective endoscopic procedures of DOAC resumption proposed in this recommendation.
considered to be at low risk of postprocedural bleeding (Table Hence, the panel was unable to make a recommendation.
3). However, we recognize that there may be a clinical concern Decisions regarding resumption of DOAC therapy should
of delayed procedural bleeding in a subgroup of patients consider the rapid onset of action, achievement of adequate
undergoing advanced endoscopic procedures. In those patients, hemostasis at the time of the procedure, the risk of delayed
decisions regarding warfarin resumption should be informed bleeding for the endoscopic procedure performed, the patient's
by achieving adequate hemostasis at the time of the procedure, risk of thrombosis, and patient preferences in consultation with
the risk of delayed bleeding associated with the endoscopic a cardiologist and hematologist.
procedure performed, the patient's risk of thrombosis, and
patient preferences in consultation with a cardiologist and Timing of P2Y12 inhibitor resumption after
hematologist. endoscopy
17. In patients who are undergoing elective endoscopic GI 18. In patients who are undergoing elective endoscopic GI
procedures whose DOAC was interrupted, we could procedures whose P2Y12 inhibitor was interrupted, we
not reach a recommendation for or against resuming could not reach a recommendation for or against re-
the DOAC on the same day of the procedure vs 1–7 suming P2Y12 inhibitor on the same day of the proced-
days after the procedure. ure vs 1–7 days after the procedure.
Summary of evidence. One prospective cohort study was Summary of evidence. Theoretically, earlier resumption of P2Y12
identified that compared the risk of bleeding based on the timing inhibitor monotherapy would tend to reduce thrombotic events
of DOAC resumption but did not compare same day with 1–7 and increase postprocedure bleeding. However, we did not
days. Radaelli et al. (96) evaluated 529 patients who interrupted identify any studies providing evidence relevant to this PICO.
DOAC therapy for an elective endoscopic procedure, including Thus, we cannot estimate the potential thrombotic or bleeding
327 with a low bleeding risk procedure and 202 with a high risk and cannot assess the balance between the desirable and
bleeding risk procedure, with 18 in the latter group receiving undesirable effects of earlier resumption.
LMWH bridging therapy, and were then followed for 30 days. Patients on P2Y12 inhibitor monotherapy are at lower cardio-
Comparing patients who resumed the DOAC on day 0–3 vs those vascular risk than those on DAPT because cardiovascular events
who resumed the DOAC after day 3, the risk of bleeding was are generally more remote in those on monotherapy. For ex-
2.3% and 11.5%, respectively (RR 0.20; 95% CI: 0.08–0.52). ample, US guidelines recommend patients remain on DAPT for
The patients receiving LMWH were not reported separately, at least 12 months after ACS before the transition to antiplatelet
and this use of bridging anticoagulation may have increased the monotherapy, with a reduction to monotherapy considered
bleeding risk in patients in whom DOAC was resumed after day after 6 months in those with high bleeding risk (54). P2Y12 in-
3. hibitor monotherapy can also substitute for ASA monotherapy
The PAUSE study was a single-arm, prospective cohort study in those with ASA hypersensitivity or GI intolerance (98,99).
that included 3,007 patients with atrial fibrillation undergoing Recent RCTs in patients with percutaneous coronary inter-
elective surgery or procedures requiring DOAC interruption vention for ACS have shown that DAPT for 1–3 months
(77). The panel was provided the raw data for the subgroup of followed by P2Y12 inhibitor monotherapy reduces major
patients undergoing a GI procedure, as described in PICO 13. bleeding with no increase in cardiovascular events compared
All GI procedures were classified as a low bleeding risk in the with continuation of DAPT for 12–24 months (55,56). Thus,
PAUSE study. The DOAC was resumed at 1.9 ± 1.5 days after patients now placed on P2Y12 inhibitor monotherapy may
have higher CV risk (because of more recent ACS), preventing results. Multicenter studies are likely required, given the low
extrapolation of baseline thrombotic risk in patients on ASA event rate of postprocedural bleeding and thrombosis.
monotherapy to those on P2Y12 inhibitor monotherapy. Finally, there is a fundamental knowledge gap in the evalu-
Similarly, we may not be able to extrapolate bleeding risk ation and characterization of GI endoscopic procedural
with ASA to P2Y12 inhibitor monotherapy. Although studies bleeding risk groups. The current estimation of procedural
have not assessed postendoscopic procedural bleeding risk, bleeding risk is highly inconsistent, derived from studies with
a meta-analysis of 5 RCTs revealed a lower risk of GIB in a serious risk of bias. A rigorous evaluation of procedural
patients taking P2Y12 inhibitor monotherapy vs ASA mono- bleeding risk with the GRADE approach will clarify and im-
therapy (OR = 0.59, 0.39–0,89) (100). prove the classification of endoscopic procedures and high-
light knowledge gaps.

FUTURE DIRECTIONS
CONFLICTS OF INTEREST
The greatest limitation to the panel's ability to provide un-
equivocal clinical recommendations was the certainty of evi- Guarantors of the article: Neena S. Abraham and Alan N.
dence in the published literature. As highlighted throughout Barkun.
this clinical practice guideline, insufficient high-quality evi- Specific author contributions: Neena S. Abraham and
dence exists in antithrombotic and antiplatelet drug users Alan N. Barkun contributed equally to this work. N.S.A.
to evaluate strategies for the temporary interruption, drug and A.N.B.: planning and execution; all authors: PICO de-
reversal, and resumption against a comparator group with velopment; G.I.L. and B.G.S.: assessment of evidence using
great certainty. In addition, we found too few studies fo- GRADE method; G.I.L. and B.G.S.: preparation of evidence
cusing on advanced endoscopic procedures to inform our profiles; N.S.A., A.N.B., L.L., J.J.T., P.A.N., J.D., and G.I.L.:
recommendations. drafting of manuscript; all authors: critical review and ap-
The GRADE approach has clearly defined criteria for proval of manuscript.
grading the certainty of evidence and the strength of re- Financial support: No external support from an industry
commendations. Accordingly, the certainty of much of the partner was obtained for this joint ACG/CAG guideline. The
relevant evidence was downgraded mainly because of indir- ACG and CAG equally shared the cost of producing this
ectness, risk of bias, and imprecision. For some clinical ques- document.
tions, we could not make a recommendation for or against Potential competing interests: N.S.A.: none declared.
the treatment strategy examined, given the very low certainty A.N.B.: Olympus Inc (Advisory Board Consulting);
or absence of evidence comparing a treatment strategy that Pendopharm Inc (research presentation); Takeda (research
is now a common clinical practice (e.g., reversal of warfarin presentation); none related to the topic of this guideline.
with vitamin K) with alternative treatment strategies. For all B.G.S.: none declared. J.D.: Potential COI related to direct
remaining clinical questions, the recommendations were con- oral anticoagulants and low-molecular-weight heparin (Pfizer,
ditional (rather than strong) because the certainty of the evi- Sanofi, Leo Pharma, Bristol-Myers Squibb, Portola, and
dence was low or very low, and the criteria for paradigmatic Janssen). L.L.: none declared. P.A.N.: Medtronic (research),
situations as described in the "METHODS" section were not Optum (advisory panel), and AliveCor (equity/royalty rela-
met. tionship); none related to the topic of this guideline. J.J.T.:
We suggest future studies focus on areas where insufficient none declared. G.I.L.: none declared.
evidence currently exists to inform clinical decisions. In par-
ticular, the potential benefit of PCC use for reversal of war-
farin in the setting of acute GIB, the appropriate timing of ACKNOWLEDGMENTS
resumption of P2Y12 receptor inhibitors and anticoagulants We gratefully acknowledge Dr. David Wan (Chair, ACG
(VKA and DOACs) after elective endoscopy, and whether it Practice Parameters Committee; ACG Practice Parameters
is necessary to interrupt P2Y12 inhibitor antiplatelet mono- Committee guideline monitor), Ms. Maria Susano and Ms.
therapy before elective endoscopy. There is also a lack of Claire Neuman (ACG), Dr. Frances Tse (Chair, CAG Practice
high certainty evidence informing optimal antithrombotic Affairs), Mr. Paul Sinclair, and Mr. Stuart Johnston (CAG) for
drug management before and after advanced endoscopic assistance with oversight on behalf of each society.
procedures.
Future observational studies hoping to influence the man-
agement of antithrombotic agents in the periendoscopic REFERENCES
period must standardize endoscopic techniques to eliminate 1. Abraham NS, Noseworthy PA, Inselman J, et al. Risk of gastrointes-
confounders for GI bleeding and ensure adequate adjustment tinal bleeding increases with combinations of antithrombotic agents
for confounders of both GI bleeding and thromboembolism. and patient Age. Clin Gastroenterol Hepatol 2020;18(2):337–46.
These studies should ensure the existence of an appropriate e319.
2. Abraham NS, Noseworthy PA, Yao X, et al. Gastrointestinal safety
comparator group and report results in sufficient detail to
of direct oral anticoagulants: A large population-based study.
allow for inclusion in future meta-analyses. Ideally, double- Gastroenterology 2017;152(5):1014–21. e1011.
blinded RCTs with adequate allocation concealment should 3. Rodríguez de Santiago E, Sánchez Aldehuelo R, Riu Pons F, et al.
be undertaken to rigorously examine the questions of tem- Endoscopy-related bleeding and thromboembolic events in pa-
porary interruption, reversal, and postprocedural resumption tients on direct oral anticoagulants or vitamin K antagonists. Clin
of antiplatelets and anticoagulants. Standardization of endo- Gastroenterol Hepatol 2021 [Epub ahead of print]. doi: 10.1016/j.
scopic techniques will increase the generalizability of RCT cgh.2020.11.037.
4. Jeon SW, Hong SJ, Lee ST, et al. Clinical practice and guidelines for 23. Khatib R, Ludwikowska M, Witt DM, et al. Vitamin K for reversal
managing antithrombotics before and after endoscopy: A national of excessive vitamin K antagonist anticoagulation: A systematic re-
survey study. Gut Liver 2020;14(3):316–22. view and meta-analysis. Blood Adv 2019;3(5):789–96.
5. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: An emerging 24. Singh S, Nautiyal A, Belk KW. Real world outcomes associated
consensus on rating quality of evidence and strength of with idarucizumab: Population-based retrospective cohort study.
recommendations. BMJ 2008;336(7650):924–6. Am J Cardiovasc Drugs 2020;20(2):161–8.
6. Spadaccini M, Albéniz E, Pohl H, et al. Prophylactic clipping after 25. Pollack CV, Reilly PA, van Ryn J, et al. Idarucizumab for dabigatran
colorectal endoscopic resection prevents bleeding of large, prox- reversal: Full cohort analysis. New Engl J Med 2017;377(5):431–
imal polyps: Meta-analysis of randomized trials. Gastroenterology 41.
2020;159(1):148–58. e111. 26. Pollack CV Jr, Reilly PA, Eikelboom J, et al. Idarucizumab for
7. Alonso-Coello P, Oxman AD, Moberg J, et al. GRADE evidence to dabigatran reversal. N Engl J Med 2015;373(6):511–20.
decision (EtD) frameworks: A systematic and transparent approach 27. Van der Wall SJ, Lopes RD, Aisenberg J, et al. Idarucizumab for
to making well informed healthcare choices. 2: Clinical practice dabigatran reversal in the management of patients with gastroin-

guidelines. Bmj 2016;353:i2089. testinal bleeding. Circulation 2019;139(6):748–56.
8. Guyatt GH, Oxman AD, Vist G, et al. GRADE guidelines: 4. Rating 28. Sewell JH, Williams L, McKnight E, et al. What is the role of
the quality of evidence: Study limitations (risk of bias). J Clin andexanet alfa in the reversal of anticoagulant effects? J Am Acad
Epidemiol 2011;64(4):407–15. PAs 2021;34(1):8–9.
9. Alonso-Coello P, Schünemann HJ, Moberg J, et al. GRADE evi- 29. Connolly SJ, Crowther M, Eikelboom JW, et al. Full study report of
dence to decision (EtD) frameworks: A systematic and transparent andexanet alfa for bleeding associated with factor Xa inhibitors. N
approach to making well informed healthcare choices. 1: Introduc- Engl J Med 2019;380(14):1326–35.
tion. BMJ 2016;353:i2016. 30. Andexxa—an antidote for apixaban and rivaroxaban. JAMA.
10. Andrews JC, Schünemann HJ, Oxman AD, et al. GRADE guidelines: 2018;320(4):399–400.
15. Going from evidence to recommendation-determinants of 31. Schulman S, Ritchie B, Nahirniak S, et al. Reversal of dabigatran-
a recommendation's direction and strength. J Clin Epidemiol associated major bleeding with activated prothrombin concentrate:
2013;66(7):726–35. A prospective cohort study. Thromb Res 2017;152:44–8.
11. Mehran R, Rao SV, Bhatt DL, et al. Standardized bleeding 32. Smythe MA, Forman MJ, Bertran EA, et al. Dabigatran versus war-
definitions for cardiovascular clinical trials: A consensus report farin major bleeding in practice: An observational comparison of
from the Bleeding Academic Research Consortium. Circulation patient characteristics, management and outcomes in atrial fibrilla-
2011;123(23):2736–47. tion patients. J Thromb Thrombolysis 2015;40(3):280–7.
12. Makris M, Greaves M, Phillips WS, et al. Emergency oral an- 33. Udayachalerm S, Rattanasiri S, Angkananard T, et al. The reversal
ticoagulant reversal: The relative efficacy of infusions of fresh of bleeding caused by new oral anticoagulants (NOACs): A sys-
frozen plasma and clotting factor concentrate on correction of the tematic review and meta-analysis. Clin Appl Thromb Hemost
coagulopathy. Thromb Haemost 1997;77(3):477–80. 2018;24(Suppl 9):117s–126s.
13. Moustafa F, Stehouwer A, Kamphuisen P, et al. Management and out- 34. da Luz LT, Marchand M, Nascimento B, et al. Efficacy and safety
come of major bleeding in patients receiving vitamin K antagonists of the drugs used to reverse direct oral anticoagulants: A systematic
for venous thromboembolism. Thromb Res 2018;171:74–80. review and meta-analysis. Transfusion 2017;57(7):1834–46.
14. Sarode R, Milling TJ Jr, Refaai MA, et al. Efficacy and safety of a 35. Costa OS, Baker WL, Roman-Morillo Y, et al. Quality evaluation
4-factor prothrombin complex concentrate in patients on vitamin K of case series describing four-factor prothrombin complex concen-
antagonists presenting with major bleeding: A randomized, plasma- trate in oral factor Xa inhibitor-associated bleeding: A systematic
controlled, phase IIIb study. Circulation 2013;128(11):1234–43. review. BMJ Open 2020;10(11):e040499.
15. Steiner T, Poli S, Griebe M, et al. Fresh frozen plasma versus 36. Zakko L, Rustagi T, Douglas M, et al. No benefit from platelet
prothrombin complex concentrate in patients with intracra- transfusion for gastrointestinal bleeding in patients taking
nial haemorrhage related to vitamin K antagonists (INCH): A antiplatelet agents. Clin Gastroenterol Hepatol 2017;15(1):46–52.
randomised trial. Lancet Neurol 2016;15(6):566–73. 37. Acosta RD, Acosta RD, Abraham NS, et al. The management of
16. Boulis NM, Bobek MP, Schmaier A, et al. Use of factor IX com- antithrombotic agents for patients undergoing GI endoscopy.
plex in warfarin-related intracranial hemorrhage. Neurosurgery Gastrointest Endosc 2016;83(1):3–16.
1999;45(5):1113–9; discussion 1118–9. 38. Baharoglu MI, Cordonnier C, Al-Shahi Salman R, et al. Platelet
17. Karaca MA, Erbil B, Ozmen MM. Use and effectiveness of pro- transfusion versus standard care after acute stroke due to sponta-
thrombin complex concentrates vs fresh frozen plasma in gastroin- neous cerebral haemorrhage associated with antiplatelet therapy
testinal hemorrhage due to warfarin usage in the ED. Am J Emerg (PATCH): A randomised, open-label, phase 3 trial. Lancet
Med 2014;32(6):660–4. 2016;387(10038):2605–13.
18. Evans G, Luddington R, Baglin T. Beriplex P/N reverses severe 39. Spiess BD, Royston D, Levy JH, et al. Platelet transfusions during
warfarin-induced overanticoagulation immediately and com- coronary artery bypass graft surgery are associated with serious
pletely in patients presenting with major bleeding. Br J Haematol adverse outcomes. Transfusion 2004;44(8):1143–8.
2001;115(4):998–1001. 40. Laine L, Barkun AN, Saltzman JR, et al. ACG clinical guideline:
19. Lubetsky A, Hoffman R, Zimlichman R, et al. Efficacy and safety of Upper gastrointestinal and ulcer bleeding. Am J Gastroenterol
a prothrombin complex concentrate (Octaplex) for rapid reversal 2021;116(5):899–917.
of oral anticoagulation. Thromb Res 2004;113(6):371–8. 41. Strate LL, Gralnek IM. ACG clinical guideline: Management
20. Vigué B, Ract C, Tremey B, et al. Ultra-rapid management of oral of patients with acute lower gastrointestinal bleeding. Am J
anticoagulant therapy-related surgical intracranial hemorrhage. In- Gastroenterol 2016;111:755.
tensive Care Med 2007;33(4):721–5. 42. Tsay C, Shung D, Stemmer Frumento K, et al. Early colonos-
21. Yasaka M, Sakata T, Naritomi H, et al. Optimal dose of prothrombin copy does not improve outcomes of patients with lower gastro-
complex concentrate for acute reversal of oral anticoagulation. intestinal bleeding: Systematic review of randomized trials. Clin
Thromb Res 2005;115(6):455–9. Gastroenterol Hepatol 2020;18(8):1696e1692–e2.
22. Holbrook A, Schulman S, Witt DM, et al. Evidence-based manage- 43. Becker RC, Scheiman J, Dauerman HL, et al. Management
ment of anticoagulant therapy: Antithrombotic therapy and preven- of platelet-directed pharmacotherapy in patients with ath-
tion of thrombosis, 9th ed: American College of Chest Physicians erosclerotic coronary artery disease undergoing elective en-
evidence-based clinical practice guidelines. Chest 2012;141(Suppl doscopic gastrointestinal procedures. J Am Coll Cardiol
2):e152S–84S. 2009;54(12):2261–76.
44. Santilli F, Rocca B, De Cristofaro R, et al. Platelet cyclooxygenase oral anticoagulants: British Society of Gastroenterology (BSG) and
inhibition by low-dose aspirin is not reflected consistently by European Society of Gastrointestinal Endoscopy (ESGE) guidelines.
platelet function assays: Implications for aspirin resistance. J Am Gut 2016;65(3):374–89.
Coll Cardiol 2009;53(8):667–77. 60. Spyropoulos AC, Brohi K, Caprini J, et al. Scientific and Stand-
45. Cheung J, Rajala J, Moroz D, et al. Acetylsalicylic acid use in ardization Committee Communication: Guidance document on
patients with acute myocardial infarction and peptic ulcer bleeding. the periprocedural management of patients on chronic oral antico-
Can J Gastroenterol 2009;23(9):619–23. agulant therapy: Recommendations for standardized reporting of
46. Sung JJ, Lau JY, Ching JY, et al. Continuation of low-dose aspirin procedural/surgical bleed risk and patient-specific thromboembolic
therapy in peptic ulcer bleeding: A randomized trial. Ann Intern risk. J Thromb Haemost 2019;17(11):1966–72.
Med 2010;152(1):1–9. 61. Douketis JD, Healey JS, Brueckmann M, et al. Perioperative bridging
47. Derogar M, Sandblom G, Lundell L, et al. Discontinuation of low- anticoagulation during dabigatran or warfarin interruption among
dose aspirin therapy after peptic ulcer bleeding increases risk of patients who had an elective surgery or procedure. Substudy of the
death and acute cardiovascular events. Clin Gastroenterol Hepatol RE-LY trial. Thromb Haemost 2015;113(3):625–32.

2013;11(1):38–42. 62. Wilke T, Bauer S, Mueller S, et al. Patient preferences for oral
48. Chan FK, Leung Ki EL, Wong GL, et al. Risks of bleeding recurrence anticoagulation therapy in atrial fibrillation: A systematic litera-
and cardiovascular events with continued aspirin use after lower gas- ture review. Patient 2017;10(1):17–37.
trointestinal hemorrhage. Gastroenterology 2016;151(2):271–7. 63. Lane DA, Meyerhoff J, Rohner U, et al. Patients' perceptions of
49. McNeil JJ, Wolfe R, Woods RL, et al. Effect of aspirin on cardiovas- atrial fibrillation, stroke risk, and oral anticoagulation treatment:
cular events and bleeding in the healthy elderly. New Engl J Med An international survey. TH Open 2018;2(3):e233–e241.
2018;379(16):1509–18. 64. Little DHW, Robertson T, Douketis J, et al. Management of
50. Bowman L, Mafham M, Stevens W, et al. ASCEND: A study of antithrombotic therapy after gastrointestinal bleeding: A mixed
cardiovascular events in diabetes: Characteristics of a randomized methods study of health-care providers. J Thromb Haemost
trial of aspirin and of omega-3 fatty acid supplementation in 2021;19(1):153–60.
15,480 people with diabetes. Am Heart J 2018;198:135–44. 65. Ara N, Iijima K, Maejima R, et al. Prospective analysis of
51. Gaziano JM, Brotons C, Coppolecchia R, et al. Use of aspirin to risk for bleeding after endoscopic biopsy without cessation of
reduce risk of initial vascular events in patients at moderate risk antithrombotics in Japan. Dig Endosc 2015;27(4):458–64.
of cardiovascular disease (ARRIVE): A randomised, double-blind, 66. Gerson LB, Michaels L, Ullah N, et al. Adverse events associ-
placebo-controlled trial. Lancet 2018;392(10152):1036–46. ated with anticoagulation therapy in the periendoscopic period.
52. Bibbins-Domingo K. Aspirin use for the primary prevention of Gastrointest Endosc 2010;71(7):1211–17. e1212.
cardiovascular disease and colorectal cancer: U.S. Preventive 67. Yanagisawa N, Nagata N, Watanabe K, et al. Post-polypectomy
services task force recommendation statement. Ann Intern Med bleeding and thromboembolism risks associated with war-
2016;164(12):836–45. farin vs direct oral anticoagulants. World J Gastroenterol
53. Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA 2018;24(14):1540–9.
guideline on the primary prevention of cardiovascular disease: A 68. Horiuchi A, Nakayama Y, Kajiyama M, et al. Removal of
report of the American College of Cardiology/American Heart As- small colorectal polyps in anticoagulated patients: A prospec-
sociation Task Force on clinical practice guidelines. Circulation tive randomized comparison of cold snare and conventional
2019;140(11):e596–e646. polypectomy. Gastrointest Endosc 2014;79(3):417–23.
54. Levine GN, Bates ER, Bittl JA, et al. 2016 ACC/AHA guideline 69. Arimoto J, Chiba H, Ashikari K, et al. Safety of cold snare
focused update on duration of dual antiplatelet therapy in patients polypectomy in patients receiving treatment with antithrombotic
with coronary artery disease: A report of the American College agents. Dig Dis Sci 2019;64(11):3247–3255.
of Cardiology/American Heart Association Task Force on clin- 70. Tafur A, Douketis J. Perioperative management of anticoagulant
ical practice guidelines: An update of the 2011 ACCF/AHA/SCAI and antiplatelet therapy. Heart 2018;104(17):1461–7.
guideline for percutaneous coronary intervention, 2011 ACCF/ 71. Douketis JD, Spyropoulos AC, Spencer FA, et al. Perioperative
AHA guideline for coronary artery bypass graft surgery, 2012 ACC/ management of antithrombotic therapy: Antithrombotic therapy
AHA/ACP/AATS/PCNA/SCAI/STS guideline for the diagnosis and and prevention of thrombosis, 9th ed: American College of Chest
management of patients with stable ischemic heart disease, 2013 Physicians evidence-based clinical practice guidelines. Chest
ACCF/AHA guideline for the management of st-elevation myo- 2012;141(Suppl 2):e326S–e350S.
cardial infarction, 2014 AHA/ACC guideline for the management 72. Douketis JD, Spyropoulos AC, Kaatz S, et al. Perioperative bridging
of patients with non-st-elevation acute coronary syndromes, and anticoagulation in patients with atrial fibrillation. N Engl J Med
2014 ACC/AHA guideline on perioperative cardiovascular evalu- 2015;373(9):823–33.
ation and management of patients undergoing noncardiac surgery. 73. Kovacs MJ, Wells PS, Anderson DR, et al. Postoperative low molec-
Circulation 2016;134(10):e123–55. ular weight heparin bridging treatment for patients at high risk of
55. Malik AH, Yandrapalli S, Shetty SS, et al. Meta-analysis of dual arterial thromboembolism (PERIOP2): Double blind randomised
antiplatelet therapy versus monotherapy with P2Y12 inhibitors in controlled trial. BMJ 2021;373:n1205.
patients after percutaneous coronary intervention. Am J Cardiol 74. Inoue T, Nishida T, Maekawa A, et al. Clinical features of post‐
2020;127:25–9. polypectomy bleeding associated with heparin bridge therapy. Dig
56. Kim BK, Hong SJ, Cho YH, et al. Effect of ticagrelor monotherapy Endosc 2014;26(2):243–9.
vs ticagrelor with aspirin on major bleeding and cardiovascular 75. Takeuchi Y, Mabe K, Shimodate Y, et al. Continuous anticoagulation
events in patients with acute coronary syndrome: The TICO and cold snare polypectomy versus heparin bridging and hot snare
randomized clinical trial. JAMA 2020;323(23):2407–16. polypectomy in patients on anticoagulants with subcentimeter polyps:
57. Boustière C, Veitch A, Vanbiervliet G, et al. Endoscopy and A randomized controlled Trial. Ann Intern Med 2019;171(4):229–37.
antiplatelet agents. European Society of Gastrointestinal endos- 76. Heublein V, Pannach S, Daschkow K, et al. Gastrointestinal endos-
copy (ESGE) guideline. Endoscopy 2011;43(5):445–61. copy in patients receiving novel direct oral anticoagulants: Results
58. Fujimoto K, Fujishiro M, Kato M, et al. Guidelines for gastroen- from the prospective dresden NOAC registry. J Gastroenterol
terological endoscopy in patients undergoing antithrombotic treat- 2018;53(2):236–46.
ment. Dig Endosc 2014;26(1):1–14. 77. Douketis JD, Spyropoulos AC, Duncan J, et al. Perioperative man-
59. Veitch AM, Vanbiervliet G, Gershlick AH, et al. Endoscopy in agement of patients with atrial fibrillation receiving a direct oral
patients on antiplatelet or anticoagulant therapy, including direct anticoagulant. JAMA Intern Med 2019;179(11):1469–78.
78. Kaatz S, Douketis JD, Zhou H, et al. Risk of stroke after surgery 91. Mantz J, Samama CM, Tubach F, et al. Impact of preopera-
in patients with and without chronic atrial fibrillation. J Thromb tive maintenance or interruption of aspirin on thrombotic and
Haemost 2010;8(5):884–90. bleeding events after elective non-cardiac surgery: The multicentre,
79. Siegal D, Yudin J, Kaatz S, et al. Periprocedural heparin bridging randomized, blinded, placebo-controlled, STRATAGEM trial. Br J
in patients receiving vitamin K antagonists: Systematic review and Anaesth 2011;107(6):899–910.
meta-analysis of bleeding and thromboembolic rates. Circulation 92. Nielsen JD, Holm-Nielsen A, Jespersen J, et al. The effect
2012;126(13):1630–9. of low-dose acetylsalicylic acid on bleeding after transu-
80. Chan FKL, Kyaw MH, Hsiang JC, et al. Risk of post-polypectomy rethral prostatectomy--a prospective, randomized, dou-
bleeding with uninterrupted clopidogrel therapy in an industry- ble-blind, placebo-controlled study. Scand J Urol Nephrol
independent, double-blind, randomized trial. Gastroenterology 2000;34(3):194–8.
2019;156(4):918–25.e1. 93. Douketis JD, Johnson JA, Turpie AG. Low-molecular-weight hep-
81. Won D, Kim JS, Ji JS, et al. Cold snare polypectomy in patients taking arin as bridging anticoagulation during interruption of warfarin:
dual antiplatelet therapy: A randomized trial of discontinuation of Assessment of a standardized periprocedural anticoagulation reg-

thienopyridines. Clin Transl Gastroenterol 2019;10(10):e00091. imen. Arch Intern Med 2004;164(12):1319–26.
82. Eisenberg MJ, Richard PR, Libersan D, et al. Safety of short-term 94. Dunn AS, Spyropoulos AC, Turpie AG. Bridging therapy in
discontinuation of antiplatelet therapy in patients with drug-eluting patients on long-term oral anticoagulants who require surgery:
stents. Circulation 2009;119(12):1634–42. The prospective peri-operative enoxaparin cohort trial (PROS-
83. Watanabe K, Nagata N, Yanagisawa N, et al. Effect of antiplatelet PECT). J Thromb Haemost 2007;5(11):2211–8.
agent number, types, and pre-endoscopic management on post- 95. Paik WH, Lee SH, Ahn DW, et al. Optimal time of resuming an-
polypectomy bleeding: Validation of endoscopy guidelines. Surg ticoagulant after endoscopic sphincterotomy in patients at risk
Endosc 2021;35(1):317–25. for thromboembolism: A retrospective cohort study. Surg Endosc
84. Whitson MJ, Dikman AE, von Althann C, et al. Is gastroduodenal 2018;32(9):3902–8.
biopsy safe in patients receiving aspirin and clopidogrel? A prospec- 96. Radaelli F, Fuccio L, Paggi S, et al. Periendoscopic management
tive, randomized study involving 630 biopsies. J Clin Gastroenterol of direct oral anticoagulants: A prospective cohort study. Gut
2011;45(3):228–33. 2019;68(6):969–76.
85. Yousfi M, Gostout CJ, Baron TH, et al. Postpolypectomy lower gas- 97. Chan FKL, Goh KL, Reddy N, et al. Management of patients on
trointestinal bleeding: Potential role of aspirin. Am J Gastroenterol antithrombotic agents undergoing emergency and elective endos-
2004;99:1785–9. copy: Joint asian pacific association of gastroenterology (APAGE)
86. Feagins LA, Iqbal R, Harford WV, et al. Low rate of postpolypectomy and asian pacific society for digestive endoscopy (APSDE) practice
bleeding among patients who continue thienopyridine therapy during guidelines. Gut 2018;67(3):405–17.
colonoscopy. Clin Gastroenterol Hepatol 2013;11(10):1325–32. 98. Jneid H, Bhatt DL, Corti R, et al. Aspirin and clopidogrel in acute
87. Cho SJ, Choi IJ, Kim CG, et al. Aspirin use and bleeding risk coronary syndromes: Therapeutic insights from the CURE study.
after endoscopic submucosal dissection in patients with gastric Arch Intern Med 2003;163(10):1145–53.
neoplasms. Endoscopy 2012;44(2):114–21. 99. Fihn SD, Gardin JM, Abrams J, et al. 2012 ACCF/AHA/ACP/
88. Dohi OUN, Gotoda T, Takizawa K, et al. Continuation of low dose AATS/PCNA/SCAI/STS guideline for the diagnosis and manage-
aspirin during per-gastric endoscopic submucosal dissection period ment of patients with stable ischemic heart disease: A report of
and increased bleeding risk : A randomized non-inferiority trial. the American College of cardiology Foundation/American Heart
Gastrointest Endoscjune 2021;93(6):AB342. Association Task Force on practice guidelines, and the American
89. Oscarsson A, Gupta A, Fredrikson M, et al. To continue or discon- College of Physicians, American association for Thoracic Surgery,
tinue aspirin in the perioperative period: A randomized, controlled Preventive Cardiovascular Nurses Association, Society for Cardi-
clinical trial. Br J Anaesth 2010;104(3):305–12. ovascular Angiography and Interventions, and Society of Thoracic
90. Antolovic D, Reissfelder C, Rakow A, et al. A randomised controlled Surgeons. Circulation 2012;126(25):e354–471.
trial to evaluate and optimize the use of antiplatelet agents in the 100. Chiarito M, Sanz-Sánchez J, Cannata F, et al. Monotherapy with
perioperative management in patients undergoing general and ab- a P2Y12 inhibitor or aspirin for secondary prevention in patients
dominal surgery: The APAP trial (ISRCTN45810007). BMC Surg with established atherosclerosis: A systematic review and meta-
2011;11:7. analysis. Lancet 2020;395(10235):1487–95.

American College of Gastroenterology Canadian Association of Gastroenterology

Uploaded by

Copyright:

Available Formats

American College of Gastroenterology Canadian Association of Gastroenterology

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

American College of Gastroenterology Canadian Association of Gastroenterology

Uploaded by

Copyright:

Available Formats

Journal of the Canadian Association of Gastroenterology, 2022, XX, 1–18

American College of Gastroenterology-Canadian

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

INTRODUCTION antiplatelet drugs such as the P2Y12 receptor inhibitors

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Management of antithrombotic agents in the setting of acute GI bleed

Vitamin K antagonist reversal

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Management of antithrombotic agents in the elective endoscopy setting

Anticoagulant interruption vs continuation

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

ASA, acetylsalicylic acid; DOAC, direct oral anticoagulant; GI, gastrointestinal.

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Table 3. Empiric endoscopic procedural bleeding risk stratification

Polypectomy (≥1 cm) EGD with/without biopsy

Risk Indication for anticoagulation

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

Downloaded from https://academic.oup.com/jcag/advance-article/doi/10.1093/jcag/gwac010/6549718 by guest on 18 March 2022

You might also like