Introduction

Anticoagulant therapy is fundamental for the prevention and treatment of thromboembolic diseases. With an aging population, the number of patients requiring long-term oral anticoagulation is increasing. Currently available options include vitamin K antagonists (VKAs) and the direct oral anticoagulants (DOACs; dabigatran, rivaroxaban, apixaban, edoxaban). A predictable pharmacology (Table 1), lower incidence of major bleeding and simplified peri-operative management are key advantages of DOACs compared with VKA. Although the overall bleeding risk in DOAC-anticoagulated patients (2.1 to 3.6% in phase III clinical trials) is lower than in VKA-treated patients,1 patients under DOACs may develop serious bleeding or need for urgent surgery. Annually, approximately 10% of patients on DOACs require invasive procedures,2 so that reversal of DOAC may also be needed for patients requiring urgent invasive procedures. DOAC-associated bleeding might impair the outcome and have detrimental or even fatal consequences. The clinical need to antagonise the anticoagulant effect depends on several factors including the anticipated bleeding risk of the procedure/operation, localisation of bleeding and the urgency for surgical intervention. Further, timing of the last intake of the anticoagulant and the renal function of the patient are the major factors influencing the elimination of the DOACs. A clinical evaluation must exclude that the invasive procedure/operation cannot be postponed to wait for the elimination of the DOAC.

Table 1 - Characteristics and indications of direct oral anticoagulants with therapeutic options for haemostasis Rivaroxaban Apixaban Edoxaban Dabigatran Drug Classification Direct FXa inhibitor Direct FXa inhibitor Direct FXa inhibitor Direct thrombin (FII) inhibitor Half-life 5 to 9 h 12 h 10 to 14 h 12 to 14 h Time to max effect 2 to 4 h 3 to 4 h 1 to 2 h 2 h Renal contraindication CrCl <15 ml min−1 CrCl <15 ml min−1 CrCl <15 ml min−1 CrCl < 30 ml min Direct antidote (dose) Andexanet alfa
Low-dose: 400 mg bolus infusion over 15 min followed by an 480 mg infusion over 2 h
High dose: 800 mg bolus infusion over 30 min followed by an 960 mg infusion over 2 h Andexanaet alfa
Low dose: 400 mg bolus infusion over 15 min followed by an 480 mg infusion over 2 h
High dose: 800 mg bolus infusion over 30 min followed by an 960 mg infusion over 2 h Andexanet alfa not approved Idarucizumab
Idarucizumab:
Infusions of 2 × 2.5 g over 5 to 10 min, infusions no more than 10 min apart. Non-specific haemostatic treatment (dose) Not approved:
PCC or aPCC at a dose of 25 to 50 IU kg−1;
rFVIIa: no recommendation Not approved:
PCC or aPCC at a dose of 25 to 50 IU kg−1;
rFVIIa: no recommendation Not approved:
PCC or aPCC at a dose of 25 to 50 IU kg−1;
rFVIIa: no recommendation Not approved:
PCC or aPCC at a dose of 25 to 50 IU kg−1;
rFVIIa: no recommendation Venous thromboembolism prophylaxis after major orthopaedic surgery (hip or knee replacement surgery)  Dosage 10 mg OD 2.5 mg BID NA 220 mg OD  Dosage adjustments No No 150 mg × 1 daily if: CrCl 30 to 50 ml min-1; or age ≥75; or concomitant use of verapamil, amiodarone or quinidine Stroke prevention in nonvalvular atrial fibrillation  Dosage 20 mg OD 5 mg BID 60 mg OD 150 mg BID  Dosage adjustments 15 mg daily if CrCl 15 to 50 ml min−1 -2.5 mg BID if 2 of 3 criteria met:
age ≥80 years;
body weight ≤60 kg; Creatinine ≥133 μmol l−
- CrCl 15 to 29 ml min−1: 2.5 mg BID 30 mg daily if:
CrCl 15 to 50 ml min-1; or - body weight ≤ 60 kg; or concomitant use of ciclosporin,
dronedarone,
erythromycin or
ketoconazole - 110 mg BID if:
age≥80 years or concomitant use of verapamil
- 110 mg or 150 mg BID if CrCl 30 to 50 ml min−1 or age 75 to 80 years Acute venous thromboembolism treatment  Dosage 15 mg BID × 21 days,
then 20 mg OD 10 mg BID x 7 days,
then 5 mg BID 60 mg OD 150 mg BID  Dosage adjustments 15 mg BID × 21 days,
then 15 mg daily if
CrCl 15 to 50 ml min−1 No dose adjustment 30 mg daily if:
- CrCl 15 to 50 ml min−1; or - body weight ≤60 kg; or concomitant use of
- ciclosporin,
- dronedarone,
- erythromycin or
- ketoconazole - 110 mg BID if age ≥80 or concomitant use of verapamil
- 110 or 150 mg BID if CrCl 30 to 50 ml min−1 or age 75 to 80 years. Extended prevention of recurrent DVT and PE  Dosage 10 mg OD
or 20 mg OD 2.5 mg BID  Dosage adjustments consider 15 mg BID instead of 20 mg BID if CrCl 15 to 50 ml min−1
10 mg: no adjustment No Acute coronary syndrome  Dosage 2.5 mg BID (plus aspirin 100 mg) NA NA NA

In principle, DOAC reversal can be achieved by specific antidotes (idarucizumab or andexanet alfa) or unspecific haemostatic agents used to enhance the haemostatic function of the coagulation system, such as prothrombin complex concentrate (PCC), activated PCC (aPCC) or recombinant factor VIIa (rFVIIa). The purpose of this guideline is to provide clinical guidance for the management of DOAC-treated patients and DOAC-induced bleeding on specific predefined clinical questions. This includes a discussion about specific reversal using direct antidotes vs. a treatment with nonspecific haemostatic agents based on current evidence.

Overview of specific antidotes and unspecific reversal agents Specific reversal: idarucizumab

Idarucizumab is a humanised, monoclonal antibody fragment against dabigatran.3 Idarucizumab binds dabigatran with high affinity (approximately 350-fold that of thrombin) neutralising the anticoagulatory effects of dabigatran. Animal models and phase I to II clinical data show that idarucizumab achieves immediate, complete and sustained biochemical reversal of dabigatran.4,5 The multicentre, prospective cohort REVERSE-AD study evaluated idarucizumab (5 g intravenously) in 503 dabigatran anticoagulated patients with major bleeding or urgent surgery. The study involved 301 patients with uncontrolled bleeding and 202 patients about to undergo an urgent procedure.6 The majority of patients had gastrointestinal and intracranial bleeding. Haemostatic efficacy was measured by the correction of the diluted thrombin time (dTT), the ecarin clotting time (ECT) and the activated partial prothrombin time (aPTT) in more than 90% of patients. Thrombotic events were observed in 4.8 and 6.8% of all patients after 30 and 90 days, respectively. Median correction of anticoagulant effect was 100% (95% confidence interval (CI), 100 to 100) based on reduction of prolonged dTT and ECT. Sixty-eight percent of patients showed cessation of bleeding within 24 h, whereas the median time to haemostasis was 2.5 h. The 30 day-mortality was 13.5 and 12.6%, respectively. Half of thrombotic events occurred within 5 days after infusion and one-third occurred after resumption of anticoagulation. Several aspects should be considered when transferring the findings of REVERSE-AD to clinical practice: it was a noncontrolled and nonrandomised study and the efficacy in terms of percentage reversal was assessed using biological markers at 4 h after the second infusion of idarucizumab. Furthermore, the duration of the haemostatic effect is to be discussed as 23% of patients had detectable dabigatran plasma levels within 12 to 24 h after administration of the antidote, which was probably because of redistribution from the extravascular to the intravascular space. This may be particularly important in patients with high plasma levels, which may be due to renal dysfunction, which, in turn, also prolongs the elimination half-life of dabigatran.7 A few case reports have reported the incomplete reversal of dabigatran following the standard dose of 5 g idarucizumab.8,9

Based on the data of the REVERSE-AD study, idarucizumab (total of 5 g iv) was approved by the European Medicines Agency (EMA) in 2016 for dabigatran reversal.

Specific reversal: andexanet alfa

Andexanet alfa is a modified, recombinant, inactive form of human FXa, with the ability to reversibly bind FXa-inhibitor molecules, thereby reducing its activity and restoring the amount of unbound endogenous FXa. Andexanet alfa therapy is administered as an intravenous (i.v.) bolus over a duration of 15 to 30 min, followed by a 2 h infusion. The ANNEXA-4 study evaluated the clinical utility of andexanet alfa as a single-group cohort study in patients who experienced acute major bleeding within 18 h after administration of apixaban (54%), rivaroxaban (40%) or enoxaparin (6%) and baseline anti-FXa activity of at least 75 ng ml−1.10 Sixty-four percent of patients suffered from ICH and 26% from gastrointestinal bleeding. Treatment with andexanet alfa resulted in a 92% reduction of anti-FXa activity. Despite a significant rebound of anti-FXa activity at 4 h after andexanet alfa infusion, in 85% of patients with gastrointestinal bleeding and in 80% of patients with intracranial bleeding, ‘excellent’ or ‘good’ haemostatic efficacy was observed at 12 h, as adjudicated according to prespecified criteria. Fourteen percent of the patients died within 30 days after study inclusion. Thrombotic events occurred in 10% of patients, none of whom had yet restarted oral anticoagulation. However, andexanet alfa binds to tissue factor pathway inhibitors (TFPI) as well, which may lead to increased thrombin generation and may explain the higher rate of thromboembolic complications compared with similar studies.11 Prior to the initiation of andexanet alfa, the dose is adjusted according to the timing of the last intake of the FXa inhibitor. Low dose consists of an intravenous bolus of 400 mg at 30 mg min−1, followed by an infusion of 480 mg at 4 mg min−1. High dose consists of a double bolus of 800 mg followed by an infusion of 960 mg at 8 mg min−1. For clinical practice, it is important to consider that the use of andexanet alfa for mild, nonlife-threatening bleeding and for patients undergoing urgent or immediate high-risk surgery has not been investigated. Similarly, to the REVERSE-AD study, the ANNEXA-4 trial was a noncontrolled and nonrandomised study with biomarkers as the primary (Reverse-AD) or co-primary (ANNEXA-4) outcome variables. Based on the data of the ANNEXA-4 trial, the European Medical Agency (EMA) and the Food and Drug Administration (FDA)-approved andexanet alfa for adult patients treated with the direct FXa inhibitors apixaban or rivaroxaban when reversal of anticoagulation is needed because of life-threatening or uncontrolled bleeding.

Haemostatic interventions with prothrombin complex, activated prothrombin complex, and rFVIIa

PCC, aPCC, and rFVIIa have been investigated for their effectiveness to treat bleeding disorders associated with recent intake of DOACs.

PCCs are lyophilised, human plasma-derived vitamin K-dependent factors containing FII (prothrombin), FVII, FIX, and FX. Potentially, three-factor PCCs that lack FVII could also be used but they are rather uncommon in Europe. As PCCs were originally used for FIX replacement in patients with haemophilia B, their potency is standardised to FIX content (about 500 international units per vial). Some PCCs might also contain protein C and S as well as small amounts of heparin or antithrombin.12

aPCC are used for controlling bleeding in haemophilia patients with inhibitors. They are composed mostly of nonactivated FII, FIX, and FX and activated FVII as well as small amounts of proteins C and S. They have been used to manage peri-operative bleeding in both nonhaemophilic and haemophilic patients. Further, PCC and aPCC have been suggested as a factor replacement approach to manage DOAC-associated life-threatening bleeding.13,14 The approach using PCC or aPCC for the management of indirect reversal strategy of DOAC-induced bleeding aims to raise the levels of vitamin K-dependent coagulation factors, notably FX for FXa inhibitors and prothrombin for dabigatran. Eventually, they will increase thrombin generation, although the exact mechanism for the treatment of DOAC-associated bleeding is not known.15 Treatment with PCC, aPCC or rFVIIa unevenly increases the concentration of several coagulation factors, including prothrombin which has the longest half-life of approximately 60 h of all these coagulation factors.16 Thrombin generation may, therefore, be enhanced for several days after the use of PCCs to treat or to prevent major bleeding in trauma or the peri-operative setting.17 This may increase the risk thromboembolic complications after treatment with PCC or aPCC.5

Recombinant FVIIa induces thrombin generation and increases FXa activity18,19 and has been associated with an increased rate of thromboembolic complications in nonhaemophilic patients with intracranial haemorrhage.20

Materials and methods

A panel of seven experts including three members of the Subcommittee ‘Fluid, Transfusion and Haemostasis’ from the European Society of Anaesthesia and Intensive Care (ESAIC) convened in 2019 to assess the latest available published evidence on the clinical management of life-threatening bleeding under DOACs. A proposal was submitted to the ESAIC guideline committee for approval of title and scope of this guideline while acknowledging the need for logistic and methodological support.

Following the 2019 Euroanaesthesia conference, scientific queries of interest were defined by the authors and formulated into 14 PICO's (population/intervention/comparison/outcome) by three authors (CvH, CFE, OG, Appendix, https://links.lww.com/EJA/A924, https://links.lww.com/EJA/A925). These PICOs were revised and merged during a discussion process among all authors resulting in a complete list of nine PICOs. These PICOs were subsequently approved by the task force in consultation with the methodologist (AA) (Table 2).

Table 2 - Summary of guidance Summary of guidance Clinical scenario: Adults under DOAC therapy undergoing urgent surgery.
Should laboratory monitoring be used in DOAC patients scheduled for urgent surgery? If yes, which laboratory monitoring should be applied? R 1.1 In patients without impaired kidney and/or liver function complying with the recommended stopping intervals (24 h for low bleeding risk, 48 to 72 h for high-bleeding risk surgery), DOAC-specific coagulation testing is not necessary. (2B) R 1.1 Measurement of DOAC levels is suggested, when stopping intervals cannot be adhered to or in patients with risk factors for elevated DOAC levels. (2C) R 1.3a Global coagulation tests including prothrombin time (PT) and activated partial thromboplastin time (aPTT) are not recommended to exclude relevant DOAC concentrations. (2B) R 1.3b In patients on FXa inhibitors with a high bleeding risk, the use of anti-FXa activity is suggested. (2B) R 1.3c In patient on dabigatran with a high bleeding risk, the use of the diluted thrombin time (dTT) or the thrombin time (TT) is suggested. (2B) Clinical scenario: Adults under DOAC therapy undergoing urgent surgery. Which test should be used: Point of care monitoring (POC) vs. non-POCT (standard laboratory) measurements and which assay (i.e. concentration/functional monitoring)? R 2.1 We do not suggest the use of nonspecific viscoelastic coagulation monitoring to reliably detect DOAC levels. (2B) R 2.2 If available, we suggest considering the use of specific tests (Ecarin clotting time and the Russell's viper venom test) in viscoelastic coagulation monitoring to exclude DOAC plasma levels. (2C) R 2.3 In the absence of specific coagulation testing, DOAC dipstick testing can be suggested to demonstrate the presence of DOACs. (2C) Clinical scenario: Adults undergoing urgent surgery with DOAC therapy. Should the prevention and/or management of DOAC induced bleeding with antidotes and nonspecific haemostatic agents (PCC, aPCC) be based on DOAC level monitoring? R 3.1 In urgent surgery and when time permits, we suggest a ‘wait and see strategy’ to reduce anticoagulant activity. A predose lab sample should be taken. (3) R 3.2a In urgent surgery, we suggest using DOAC concentration measurement to guide the administration of antidotes or nonspecific haemostatic agents. (2C) R 3.2b If DOAC monitoring is not available, and surgery cannot be delayed, antidote and nonspecific haemostatic treatment should depend on the clinical severity of bleeding. (3) R 3.3a When urgent surgery with high risk of bleeding cannot be delayed, and if relevant residual concentration of dabigatran is suspected, specific antidote therapy with idarucizumab is recommended without waiting for DOAC level monitoring. However, a predose lab sample should be taken. (1C) R 3.3b If idarucizumab is not available, PCC or aPCC may be used for the urgent surgical setting in patients on dabigatran without waiting for DOAC level monitoring. A predose lab sample should be taken. (3) R 3.4 When urgent surgery with high risk of bleeding cannot be delayed, and if relevant residual concentration of FXa inhibitors is suspected, andexanet alfa, PCC or aPCC is suggested without waiting for DOAC level monitoring. However, a predose lab sample should be taken. (3) R 3.5 In urgent surgery with a high risk of bleeding, the plasma concentrations of DOACs above 50 ng ml−1 may be considered for haemostatic or antidote intervention. (3) R 3.6 In cardiac surgery under FXa inhibitors, we recommend not to use andexanet alfa. The use of haemadsorption filters may be considered. (3) Clinical scenario: Adults undergoing urgent surgery with DOACs therapy. Should laboratory measurements be performed after reversal (which time frame of measurements)? R 4.1 After specific reversal of dabigatran with idarucizumab, we suggest to assess dabigatran concentrations by the diluted thrombin time (dTT) test or the thrombin time (TT) regularly for at least 48 h because of potential drug rebound. (2B) R 4.2 After specific reversal of direct FXa inhibitors with andexanet alfa, caution is advised in interpretation of the concentration measurements as anti-FXa activities are influenced by andexanet alfa. (3) R 4.3 After administration of nonspecific haemostatic treatment in patients with elevated or suspected high levels of direct FXa inhibitors, it is unclear when and whether anti-FXa levels should be re-assessed. Conventional coagulation testing including PT or aPTT may indicate normalisation for several hours despite insufficient haemostasis. (3) Clinical scenario: DOAC-treated adult patients with traumatic and nontraumatic intra cerebral haemorrhage without need for surgery. Are antidotes or nonspecific haemostatic agents indicated for DOAC-treated patients with traumatic and nontraumatic ICH without need for surgery? R 5.1 We recommend antidote reversal or nonspecific haemostatic agents to prevent increasing haematoma volume. (1C) R 5.2a We recommend the use of idarucizumab for the reversal of dabigatran-associated intracerebral bleeding. (1C) R 5.2b PCC or aPCC may be considered for patients taking dabigatran if idarucizumab is not available. (3) R 5.3 We suggest the use of andexanet alfa or PCC to prevent increasing haematoma volume following apixaban and rivaroxaban-associated intracerebral bleeding. If andexanet alfa or PCC are not available, aPCC may be considered. (2C) R 5.4 PCC may be considered for patients taking edoxaban. (3) Clinical Scenario: Nonbleeding adults with overdose of DOACs not considered for urgent or elective surgery. Should reversal agents be used to manage nonbleeding patients with an overdose of DOAC? R 6.1 We suggest not to reverse dabigatran or FXa inhibitors in the absence of bleeding. (3). R 6.2 We suggest general measures to eliminate FXa inhibitors, which may include stimulation of diuresis and/or haemadsorption. (2 C) R 6.3 We suggest the stimulation of diuresis and the use of haemodialysis in the haemodynamically stable patient with dabigatran overdose. In early dabigatran overdose, activated charcoal may be considered. (2C) Clinical scenario: Adult patients under FXa inhibitor therapy, who present with severe bleeding in urgent surgical or nonsurgical settings. Should andexanet alfa or PCC, aPCC or rFVIIa be used to manage FXa inhibitor-associated bleeding in urgent surgical or nonsurgical settings? R 7.1 We recommend that PCC or andexanet alfa should be considered in patients under FXa inhibitor therapy presenting with severe bleeding. However, the superiority of one agent over another has not been demonstrated. (1C) R 7.2 In the absence of the availability of andexanet alfa and PCC, aPCC may be considered in patients under FXa inhibitor therapy presenting with severe bleeding. (2C) R 7.3 Due to the paucity of clinical data, we are unable to provide any recommendation for the use of rFVIIa in patients under FXa inhibitor therapy presenting with severe bleeding. (3) Clinical scenario: Adult patients under dabigatran therapy, who present with severe bleeding in urgent surgical or nonsurgical settings. Should idarucizumab or PCC, aPCC or rFVIIa be used to manage dabigatran associated bleeding in urgent surgical or nonsurgical settings? R 8.1 We recommend that idarucizumab should be considered in patients under dabigatran therapy presenting with severe bleeding or in urgent surgical or nonsurgical settings. (1 C) R 8.2 In the absence of the availability of idarucizumab, we suggest the use of PCC or aPCC. However, the superiority of one agent over another has not been demonstrated. (2C) R 8.3 Due to the paucity of clinical data, we are unable to provide any recommendation for the use of rFVIIa. (3) Clinical scenario: Invasive nonsurgical procedures with a high risk of bleeding under DOAC therapy in adults. Should reversal agents be used before an urgent invasive procedure, including regional anaesthesia, aortic stent placement, and so forth? R 9.1 In patients on dabigatran who are undergoing urgent invasive procedures with a high risk of bleeding, idarucizumab is recommended to reduce levels of dabigatran in order to normalise coagulation. (1C) R 9.2 Andexanet alfa has not been investigated before urgent invasive procedures. We are unable to provide any recommendation for the use of andexanet alfa nor for any haemostatic agents. (3)

aPCC, activated PCC; aPTT, activated partial thromboplastin time; dTT, diluted thrombin time; PCC, prothrombin complex concentrate; PT, prothrombin time; rFVIIa, recombinant activated factor VII; TT, thrombin time.

Criteria for considering studies for data analysis and search results

Data analysis was based on all randomised, parallel, quasi-randomised studies (including cross-over design) and observational studies that addressed the above queries. Systematic reviews and meta-analyses were considered on a case-by-case basis when meeting inclusion criteria. Data from quasi-randomised and observational and large retrospective studies were included as very few if any randomised controlled trials (RCTs) were anticipated. Case reports were removed from all databases. However, in case of a lack of evidence, case reports were allowed to be added by the individual authors to answer the PICOs.

Types of interventions and comparators

Searches were based on DOAC (Population) and the intervention (Measurement of DOAC/anticoagulants or reversal/antidote).

Search methods for identification of studies

The electronic database search was run on 10 February 2021 by Cochrane Trial Search Specialist (JV) and included articles published since 2010 to increase clinical relevance. For more information on the search strategy and the search results, please see appendix, https://links.lww.com/EJA/A924, https://links.lww.com/EJA/A925. The panel members were also encouraged to add any missing papers of interest that they were aware of and to conduct related searches themselves. This was carried out by all authors subsequently.

Search results

Literature search bundles and corresponding structured search strategies were developed and conducted in Medline (3890), EMBASE (3830), Web of Science (2828), Epistemonikos (77) and Cochrane Central Register of Controlled Trials (CENTRAL, 769). After removing all duplicates, the titles resulting from the searches were screened by all authors. Relevant full-text articles were retrieved and assessed in detail. Furthermore, the authors were allowed to recommend inclusion of any landmark study or relevant articles published after the search date during the entire process of guideline creation. The decision to include any relevant study was undertaken subsequent to voting by the entire task force for each article of relevance. This decision was taken to compensate for the COVID-related delay in the guideline drafting process.

The authors aimed to include a restricted number of supporting references to support each recommendation as part of a brief accompanying rationale, with studies of the best available quality from any publication date given priority. A total number of 138 studies were included for this evidence synthesis.

Recommendations were formulated and graded according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system (Table 3)21 and in accordance with ESAIC methodology guidelines. Recommendations, grading and rationales were initially drafted and critically reviewed by the entire author group for review prior to the virtual consensus process and in between each round of virtual meetings. The quality of the literature cited to support each recommendation was reviewed separately by the methodologist (AA), who applied Cochrane risk-of-bias assessment criteria for RCTs. For non-RCT, the Scottish Intercollegiate Guideline Network (SIGN) checklists (https://www.sign.ac.uk/what-we-do/methodology/checklists/) was used to assess risk of bias of the included studies prior to grading and drafting recommendations, suggestions or clinical practice statements in accordance with GRADE for guidelines and ESAIC SOPs for guidelines.

Table 3 - GRADE definitions Grade of Recommendation Clarity of risk/benefit Quality of supporting evidence 1A
Strong recommendation, high quality evidence Benefits clearly outweigh risk and burdens, or vice versa. Consistent evidence from well performed randomised, controlled trials or over-whelming evidence of some other form. Further research is unlikely to change our confidence in the estimate of benefit and risk. 1B
Strong recommendation, moderate quality evidence Benefits clearly outweigh risk and burdens, or vice versa. Evidence from randomised, controlled trials with important limitations (inconsistent results, methodologic flaws, indirect or imprecise), or very strong evidence of some other research design. Further research (if performed) is likely to have an impact on our confidence in the estimate of benefit and risk and may change the estimate. 1C
Strong recommendation, low quality evidence Benefits appear to outweigh risk and burdens, or vice versa. Evidence from observational studies, unsystematic clinical experience, or from randomised, controlled trials with serious flaws. Any estimate of effect is uncertain. 2A
Weak recommendation = suggestion, high quality evidence Benefits closely balanced with risks and burdens. Consistent evidence from well performed randomised, controlled trials or overwhelming evidence of some other form. Further research is unlikely to change our confidence in the estimate of benefit and risk. 2B
Weak recommendation = suggestion, moderate quality evidence Benefits closely balanced with risks and burdens, some uncertainly in the estimates of benefits, risks and burdens. Evidence from randomised, controlled trials with important limitations (inconsistent results, methodologic flaws, indirect or imprecise), or very strong evidence of some other research design. Further research (if performed) is likely to have an impact on our confidence in the estimate of benefit and risk and may change the estimate. 2C
Weak recommendation = suggestion, low quality evidence Uncertainty in the estimates of benefits, risks, and burdens; benefits may be closely balanced with risks and burdens. Evidence from observational studies, unsystematic clinical experience, or from randomised, controlled trials with serious flaws. Any estimate of effect is uncertain. 3
Clinical practice statement High uncertainty in the estimates of benefits, risks, and burdens; benefits may outweigh risks and burdens. Evidence from observational studies, unsystematic clinical experience, case reports or extrapolated from other settings and populations from trials with serious flaws. Any estimate of effect is uncertain.

Authors participated in a series of six virtual consensus conferences during 2022, and three virtual Delphi method rounds in 2023 during which the wording and grading of each recommendation was finalised and confirmed by voting members of the expert panel (Table 2). For the clinical scenarios of urgent surgery, life-threatening and non-life-threatening bleeding different flow charts comprising relevant recommendations and clinical practice statements have been developed that may help to inform clinicians to manage these scenarios successfully (Figs. 1–3).

Fig. 1:

Treatment algorithm for the management of patients with confirmed intake of direct oral anticoagulants before urgent surgery.

Fig. 1(Continued):

Treatment algorithm for the management of patients with confirmed intake of direct oral anticoagulants before urgent surgery.

Fig. 2:

Treatment algorithm for the management of patients with non-life threatening bleeding.

Fig. 2 (Continued).:

Treatment algorithm for the management of patients with non-life threatening bleeding.

Fig. 3:

Treatment algorithm for the management of severe (haemodynamically unstable) or life-threatening bleeding (intracerebral, epidural, intraspinal, gastrointestinal, traumatic or other refractory bleeds).

Fig. 3(Continued).:

Treatment algorithm for the management of severe (haemodynamically unstable) or life-threatening bleeding (intracerebral, epidural, intraspinal, gastrointestinal, traumatic or other refractory bleeds).

Grading was confirmed and disagreements resolved in consultation with the methodologist (AA). During the virtual consensus conferences, it became apparent to the entire task force that there was a great degree of overlap between the various PICOs and the included references. This resulted initially in many similar recommendations and suggestions for various PICOs. To increase the clinical relevance, while avoiding repetition and with the aim of a higher degree of implementation in clinical practice, the task force decided after consultation with the methodologist to merge several overlapping PICOs into nine final PICOs prior to the Delphi method rounds:

1. Should laboratory monitoring be used in DOAC patients scheduled for urgent surgery? If yes, which laboratory monitoring should be applied? 2. Which test should be used: point of care monitoring vs. non-POCT (standard laboratory) measurements and which assay (i.e. concentration/functional monitoring)? 3. Should the prevention and/or management of DOAC-induced bleeding with antidotes and nonspecific haemostatic agents (PCC, aPCC) be based on DOAC level monitoring? 4. Should laboratory measurements be performed after reversal (which time frame of measurements)? 5. Are antidotes or nonspecific haemostatic agents indicated for DOAC-treated patients with traumatic and nontraumatic ICH without need for surgery? 6. Should reversal agents be used to manage nonbleeding patients with an overdose of DOAC? 7. Should PCC, andexanet alfa, aPCC or rFVIIa be used to manage factor Xa inhibitor-associated bleeding? 8. Adult patients under dabigatran therapy, who present with severe bleeding in urgent surgical or nonsurgical settings. 9. Invasive nonsurgical procedures with a high risk of bleeding under DOAC therapy in adults.

For more information on which PICOs were merged, please see Supplementary Appendix, https://links.lww.com/EJA/A924, https://links.lww.com/EJA/A925. Following final revisions, manuscript collation and approval by the author group, the manuscript was submitted to ESAIC guideline committee, ESAIC board and all ESAIC members for an open-peer review assessment in May 2023.

Discussion

The present guideline summarises available evidence on proper management of bleeding complications with or without the need of acute surgery among DOAC-treated patients. The original purpose of this project was to provide clinicians with a set of meaningful recommendations regarding monitoring, strategies for planning surgery and choice of antidote or haemostatic therapy.

In summary, our recommendations are clearly dependent on the specific clinical situation. As all DOAC have a relatively rapid effect offset, a ‘wait-and-see strategy’ is in general recommended awaiting natural elimination of the drug. However, in life-threatening bleeding, or if emergency surgery is needed, active treatment with antidote or haemostatic therapy may be needed. In all cases, laboratory monitoring is recommended, although in urgent situations, results should not be awaited.

The era of specific antidotes for DOAC has evolved within a short period of time, whereby the number and scientific quality of relevant studies are low, especially lacking published randomised controlled trials. This implies that only a few grade 1 recommendations could be given and hampers drawing definite conclusions about the efficacy of specific reversal and nonspecific haemostatic agents.

For life-threatening or nonlife threatening bleeding in which a ‘wait and see strategy’ is not clinically applicable, we recommend the antidote idarucizumab for dabigatran, PCC for edoxaban while the antidote andexanet alfa and PCC are equal treatment options for apixaban and rivaroxaban. This aligns with recent published guidelines from the European Society of Anaesthesiology and Intensive Care from April 2023,22 with the exception that PCC rather than andexanet alpha in patients treated with anti-FXa agents (rivaroxaban and apixaban) is recommended (grade 2C) in bleeding patients and opposite to a European guideline on management of major bleeding and coagulopathy following trauma that recommends only PCC if andexanet alfa is unavailable in life-threatening bleeding and the presence of apixaban or rivaroxaban.23

During the reviewing process of this guideline preliminary results of the prospective, randomised ANNEXA-I study that investigated the efficacy of andexanet alfa vs. usual care therapy in patients with intracranial haemorrhage were published during an oral presentation at the World Stroke Congress in October 2023. The study was prematurely stopped after a planned interim analysis of 452 patients and the results indicate a significantly higher efficacy of andexanet alfa as compared with usual care that including 86% of patients having been treated with PCC. However, the treatment with andexanet alfa showed a significantly higher rate of thromboembolic events and the mortality between the groups was not different at 30 days. The majority of the guideline authors decided to not wait for the full publication of this study, as this would unnecessaril