Perioperative Management of Anticoagulant and Antiplatelet Therapy
The management of patients who are receiving an anticoagulant or antiplatelet drug and require surgery or an invasive procedure is a common clinical problem encountered by a broad spectrum of health care professionals. This review provides an evidence-based but practical approach to common clinical scenarios involving patients who require an elective surgery/procedure and are receiving either a vitamin K antagonist, a direct oral anticoagulant, or single or dual antiplatelet therapy. This review also addresses the role of perioperative heparin bridging and the management of patients who are receiving an anticoagulant drug and need urgent surger
Introduction
The management of patients who are receiving an anticoagulant or antiplatelet drug and require surgery or an invasive procedure is a common clinical problem1,2 that is increasing due to an aging population, with more patients likely to require a surgery/procedure.3-5
This review aims to inform the clinician managing patients who require surgery or a procedure and are receiving a vitamin K antagonist (VKA) and may need heparin bridging, a direct oral anticoagulant (DOAC), or single or dual antiplatelet therapy with acetylsalicylic acid (ASA) and/or a P2Y12 inhibitor
Foundational Concepts in Perioperative Antithrombotic Management
For patients who need an elective surgery/procedure, perioperative antithrombotic management is anchored on an assessment of patient- and procedure-related risks for bleeding and thromboembolism (Tables 1 and 2), an understanding of pharmacologic properties of antithrombotic drugs pertinent to perioperative management (Tables S1 and S2 in the Supplementary Appendix), and ensuring adequate communication and management consensus among the health care team
Risk assessment for bleeding and thromboembolism
Given the heterogeneity in thromboembolic and bleed risk profiles of individual patients and various surgeries/ procedures,6-8 perioperative antithrombotic management adopts a patient-centric approach. Empiric risk classification schemes for bleed risk (Table 1) and thromboembolic risk (Table 2) 1,9,10 have been developed to guide management but may be overridden depending on individual patient circumstances and clinical judgment. In patients with coronary artery disease, risk classification schemes are provided elsewhere but are also empiric.11-13
Pharmacologic properties of anticoagulant and antiplatelet drugs
Knowledge of key pharmacokinetic and pharmacodynamic properties of anticoagulant and antiplatelet drugs, as shown in Tables S1 and S2, respectively, in the Supplementary Appendix, is important to minimize bleeding and thromboembolism risks.14,15 The timing of preoperative interruption of VKAs and DOACs is informed by the anticoagulant t1/2 (elimination half-life), whereas the timing of postoperative anticoagulant resumption is informed by the tmax (time to maximum anticoagulant activity).16 These properties also guide the use of heparin bridging, usually with low-molecular-weight heparin (LMWH).17 Antiplatelet potency and platelet function reversibility also inform perioperative drug interruption and resumption.1,18
Patients Taking a VKA
Although VKAs comprise warfarin, acenocoumarol, and phenprocoumon, most evidence around perioperative management pertains to warfarin
Perioperative VKA continuation
There are several surgeries/procedures classified as having minimal bleed risk (Table 1) in which VKAs can be continued without interruption.19-25 Evidence from small, randomized trials and cohort studies supports the safety of continuing VKAs around minor dental and dermatologic procedures and cataract surgery, which is a mainly avascular procedure.1,9,10 There is emerging evidence as to the safety of continuing VKAs around gastrointestinal endoscopic procedures when done for screening purposes that require biopsies but where polypectomy is not anticipated.26 This approach is supported by practice guidelines but does not extend to DOACs due to lack of evidence for periendoscopic continuation. Table 1 classifies bleed risk in common gastrointestinal procedures, which guides whether anticoagulants are interrupted or continued. Well-designed randomized trials, including BRUISE CONTROL-1,27 have shown that in patients with cardiac device implantation, a warfarin continuation approach is associated with less bleeding than warfarin interruption with heparin bridging.22,28 Preventing device pocket-site bleeding is important because its occurrence increases the risk for infection, which can have major clinical consequences.29,30 In patients in whom VKAs are continued, it is assumed that the degree of anticoagulation, as measured by the international normalized ratio (INR), will not be excessive, with an INR less than 3.0 at the time of the procedure. In certain dental or dermatologic procedures, patients may require adjunctive hemostatic measures such as oral tranexamic mouthwash or pressure dressings. For cataract surgery, it is assumed (as is the current practice in most patients) that topical instead of retrobulbar anesthesia will be used because the latter may increase the risk of retro-orbital bleeding.
Perioperative heparin bridging
In a surgery/procedure that requires VKA interruption, heparin bridging may be considered, typically with a therapeutic-dose LMWH regimen (e.g., dalteparin, 100 IU/kg twice daily), which is the most widely studied and which usually (but not always) is administered before and after a surgery/procedure. The therapeutic premise of heparin bridging has been to shorten the period before and after a surgery/procedure so that patients are not fully anticoagulated while a VKA is interrupted and, in turn, to reduce the risk for perioperative stroke and other thromboembolism. The perioperative milieu, however, is one in which thromboembolism may occur through various mechanisms that would not be affected by perioperative heparin bridging.7 Large randomized controlled trials, including BRIDGE and PERIOP-2, and multiple observational studies provide consistent evidence that administering LMWH bridging during warfarin interruption in patients with atrial fibrillation or mechanical heart valves does not affect the incidence of perioperative thromboembolism but increases the risk for major bleeding by threefold or higher.31-34 Moreover, the occurrence of perioperative bleeding, although typically treatable, has the consequence of delaying the resumption of anticoagulant therapy postprocedure, which then exposes patients to a downstream increased risk for thromboembolism.35 Recent practice guidelines recommend against routine heparin bridging in VKA-treated patients with atrial fibrillation and advise against heparin bridging in patients with venous thromboembolism and in selected patients with a mechanical heart valve.1
Selected patients in whom heparin bridging may be warranted include those categorized as at high risk for thromboembolism (Table 2). The approach to warfarin interruption and bridging with an LMWH (Fig. 1) aims to minimize both the time patients are not fully anticoagulated and the bleeding risk. The rapid peak effect of LMWHs, occurring 3 to 4 hours after administration, prompts the need for flexibility in the timing of postoperative resumption: approximately 24 hours for low/moderate-bleed-risk and 48 to 72 hours for high-bleed-risk surgery/procedures.36 With selected high-bleed-risk surgeries (e.g., cardiac or intracranial), the risks and consequences of bleeding may preclude the use of postoperative bridging with therapeutic-dose LMWH, and in such patients, a prophylactic (low-dose) LWMH regimen (e.g., dalteparin, 5000 IU daily) can be given for 3 to 5 days while the VKA is resumed. Avoidance of postoperative heparin bridging is also advised in patients when surgical hemostasis is inadequate. Finally, routinely measuring antifactor Xa levels during perioperative bridging has not been studied31-33 and is unlikely to have clinical utility.
Timing of VKA interruption and resumption
Stopping warfarin (36- to 42-hour half-life) 5 days before the day of the surgery/procedure is sufficient to ensure that more than 90% of such patients will have a normal (<1.3) or near-normal (1.3 to 1.4) INR at the time of the surgery, thereby obviating the need for routine preoperative INR testing.35,37 Shorter (2 to 3 days) and longer (7 to 10 days) interruptions are needed for acenocoumarol and phenprocoumon, respectively. Warfarin can be resumed within 24 hours of, and often on the evening of, the surgery/procedure. Although doubling the first postoperative dose of warfarin, as compared with resuming the usual warfarin dose, has been supported by some studies,38 the benefit in terms of time to reach a therapeutic INR is limited (~1 day faster).35,39,40 An interruption of longer than 5 days may be required in patients who have delayed warfarin elimination due to advanced age or interacting drugs that interfere with warfarin metabolism and in patients with a recent high (>3.5) INR or with a higher target INR (i.e., 2.5 to 3.5).41,42
Patients Taking a DOAC
In patients taking DOACs (apixaban, dabigatran, edoxaban, or rivaroxaban), perioperative management is anchored on their 10- to 14-hour elimination half-lives (18 to 24 hours for dabigatran in patients with impaired renal function), which informs preoperative interruption, and their rapid onset of action that occurs 2 to 3 hours after intake, which informs postoperative resumption.43 Key pharmacokinetic properties of DOACs relevant for perioperative management are shown in Table S1. This rapid offset and onset of action obviate the need for heparin bridging. Subanalyses and a meta-analysis of randomized trials comparing DOAC versus warfarin therapy in patients with atrial fibrillation who required an elective surgery found that patients who received heparin bridging during DOAC interruption had an approximately threefold higher risk of major bleeding with no effect on thromboembolism.44-47
Perioperative DOAC continuation
Most studies assessing DOACs involve patients with a cardiac device implant. These studies suggest that DOACs can be safely continued around pacemaker or internal cardiac defibrillator implantation48 and in patients having an atrioventricular node ablation procedure.49-51 A randomized trial comparing the continuation of dabigatran versus 1 to 2 days of interruption found no significant difference in rates of pocket hematoma (2.1% vs. 2.1%).52 A subanalysis of randomized trials comparing DOACs versus warfarin use in patients with atrial fibrillation found less bleeding with periprocedural continuation of DOACs than with warfarin continuation.47 For minor dental and skin procedures and cataract surgery, DOAC continuation appears to be safe,53-55 preferably with a delayed dose (for once-daily DOACs) until after a procedure (or omitting the morning dose for twice-daily DOACs) to avoid a peak anticoagulant effect during the procedure.
DOAC interruption and resumption
In patients in whom DOAC interruption is required, evidence for recommended management (Fig. 2) is derived mainly from the PAUSE trial, in which 3007 patients with atrial fibrillation who were taking a DOAC (1257 on apixaban, 668 on dabigatran, and 1082 on rivaroxaban) received standardized perioperative management based on pharmacokinetic principles.56 This approach was associated with low 30-day perioperative rates of thromboembolism (0.16% to 0.60%) and major bleeding (0.9% to 1.85%), depending on the DOAC.
In patients having a low/moderate-bleed-risk surgery/ procedure, which encompasses most day surgeries that typically last less than 1 hour and most ambulatory procedures, PAUSE-based management requires 1 day off the DOAC before the surgery/procedure, which corresponds to a 30- to 36-hour interruption interval (or two to three DOAC half-lives) between the last dose and the procedure. In patients having a high-bleed-risk surgery/ procedure, which comprises most surgeries lasting 1 hour or more or any neuraxial (epidural or spinal) anesthesia, PAUSE-based management requires 2 days off the DOAC before the surgery/procedure, which corresponds to a 60- to 68-hour interruption interval (or four to five DOAC half-lives) between the last dose and the procedure. In high-bleed-risk patients, this management was associated with rates of major bleeding of 0.9% to 3.0%, depending on the DOAC patients were receiving. Although laboratory testing did not inform DOAC interruption, 98.8% of such patients had a DOAC level less than 50 ng/ml at the time of surgery, and depending on the DOAC 86% to 98% had a DOAC level less than 30 ng/ml. Another study of nonstandardized perioperative DOAC management in 422 patients found that a 48- to 72-hour interruption interval was optimal, associated with 95% of patients having a preoperative DOAC level of 30 ng/ml or less,57,58 thereby supporting the PAUSEbased management.
After the surgery/procedure, DOAC resumption is flexible to allow variability in surgery/procedure site hemostasis, whereby DOACs can be resumed approximately 24 hours after a low/moderate-bleed-risk procedure and approximately 48 to 72 hours after a high-bleed-risk procedure. In cases where postoperative resumption of oral medications is delayed, prophylactic (low-dose) LMWH can be given in surgeries associated with a high venous thromboembolism (VTE) risk.1 Of note, some anesthesia societies recommend an empiric-based DOAC interruption interval 3 days or more or 72 hours before neuraxial (spinal/epidural) anesthesia or a procedure (e.g., nerve root block),59,60 which is slightly longer than the 60- to 68-hour interval assessed in PAUSE We suggest preoperative consultation with the treating anesthesiologist if there is uncertainty as to DOAC interruption.
Preoperative measurement of DOAC levels
To measure the anticoagulant effect of DOACs, antifactor Xa levels can be used for apixaban, edoxaban, and rivaroxaban, and the dilute thrombin time for dabigatran.61-63 However, these tests are not widely and rapidly accessible, and even if the tests are available there is no consensus as to the DOAC level that would enable a surgery/procedure to proceed safely.64,65 Some have advocated that a level less than 50 ng/ml is a safe threshold to allow surgery to proceed, whereas others have suggested a level less than 30 ng/ml, which is the limit of DOAC level detection of some assays.66,67 A subanalysis of the PAUSE study suggests that DOAC levels less than 30 ng/ml or 30 to 50 ng/ml are not associated with an increased risk for perioperative bleeding.68,69 Based on this uncertainty, DOAC level testing is not recommended for routine use perioperatively in patients having an elective surgery/procedure.1
Patients Receiving Antiplatelet Therapy
Antiplatelet drugs, comprising ASA and the P2Y12 inhibitors clopidogrel and prasugrel, irreversibly inhibit platelet function so that 7 to 10 days (i.e., platelet lifespan) of preoperative interruption is needed to fully restore platelet function,18 whereas with the P2Y12 inhibitor ticagrelor, which reversibly inhibits platelet function, 2 to 4 days of interruption are needed to restore platelet function (Tables S1 and S2).70 With postoperative management, a maximal antiplatelet effect occurs within minutes after resuming ASA, within 2 hours after resuming ticagrelor, at approximately 3 days after resuming prasugrel, and at 4 to 5 days after resuming clopidogrel at a 75-mg maintenance dose.71-73 An overall approach to perioperative antiplatelet management is shown in Figure 3.
Noncardiac surgery
ASA interruption may be favored in patients with stable coronary artery disease who have not had prior coronary revascularization and are undergoing a high-bleed-risk surgery, whereas ASA continuation may be favored in patients undergoing vascular procedures (e.g., carotid endarterectomy) or in those with a recent (1 year or less) myocardial infarction or with prior coronary revascularization. Randomized trials have assessed whether to continue or interrupt ASA for noncardiac surgery.74 These studies, of which the PEP and POISE-2 trials are the largest, suggest that continuing ASA does not appear to reduce the incidence of myocardial ischemic events or overall mortality but confers an approximately 1% increased risk for major bleeding.75,76 A limitation of these trials was that there was no comparison, except in subgroups, of ASA continuation versus 7- to 10-day ASA interruption that would reflect management options in clinical practice. In POISE-2, the increase in bleeding with perioperative ASA use occurred in patients who were prior nonusers and initiated ASA before surgery but not in ASA users who continued ASA perioperatively. Overall, perioperative ASA use is associated with a decrease in the risk for major (venous and arterial) thromboembolism (relative risk [RR], 0.74; 95% confidence interval [CI], 0.58 to 0.94) while increasing the risk for major bleeding (RR, 1.31; 95% CI, 1.15 to 1.50).77 Practice guidelines suggest that nonusers of ASA should not initiate ASA perioperatively, whereas users of ASA should continue ASA perioperatively.1,78,79 An important caveat mentioned in these guidelines is that perioperative ASA interruption can be justified in many patients, such as those undergoing a high-bleed-risk surger
Before Coronary Artery Bypass Grafting Surgery
Large observational studies have supported perioperative ASA use around coronary artery bypass grafting (CABG) surgery by demonstrating a reduction in cardiovascular events without increasing bleeding.80 In the ATACAS trial, patients received ASA (100mg) starting 1 to 2 hours before CABG (prior users stopped ASA 4 days before surgery) or placebo, with resumption within 24 hours after CABG.81 The trial found no significant effect of ASA use on 30-day postoperative rates of myocardial infarction (13.8% vs. 15.8%; RR, 0.87; 95% CI, 0.71 to 1.07), death, or other cardiovascular outcomes, and ASA use did not increase the risk for reoperation related to bleeding (1.8% vs. 2.1%; RR, 0.87; 95% CI, 0.47 to 1.6). The generalizability of results may be questioned because interrupting ASA and restarting it 1 to 2 hours before CABG (representing the ASA continuation group) is not a widespread practice. Among patients who are receiving ASA and a P2Y12 inhibitor, subanalyses of randomized trials involving patients with acute coronary syndromes, of whom a subgroup required CABG surgery, suggested that interrupting clopidogrel or ticagrelor approximately 5 days before surgery minimized bleeding and did not affect cardiovascular event risk. Practice guidelines generally suggest interrupting the P2Y12 inhibitor at 3 to 5 days (ticagrelor), 5 days (clopidogrel), or 7 to 10 days (prasugrel) before CABG surgery.1,13
In patients who developed an acute coronary syndrome, were treated with percutaneous coronary intervention and dual antiplatelet therapy, and require urgent CABG surgery, one management option is to wait at 3 to 5 days to allow the antiplatelet effect of the P2Y12 inhibitor to recede sufficiently. If urgent CABG surgery is needed, transfusing platelets with the intent of providing fully functional platelets may seem sensible,82 but this approach may only provide hemostatic benefit if antiplatelet drugs are not circulating in plasma, leading to inhibition of newly transfused platelets.83,84
Patients With Coronary Stents
About 10% to 15% of patients with coronary stents will require surgery within 2 years of stenting.85,86 Perioperative management can vary depending on the timing of stent placement, the stent type (drug-eluting or bare metal), the stent location, whether the stent is in a dominant or nondominant coronary artery and the number and length of stents.87,88 Management options vary from stopping both antiplatelet drugs and bridging with a glycoprotein IIb/IIIa inhibitor or cangrelor to continuing both drugs without interruption.89,90 The interval between coronary stenting and the surgery will also affect management because the risk for cardiovascular events is highest within 4 to 6 weeks after stenting but may persist for 6 to 12 months.91,92 In a subanalysis of the POISE-2 trial that assessed 470 patients with a coronary stent who required major noncardiac surgery,93,94 those who continued ASA were at lower risk for developing myocardial infarction than patients who did not receive perioperative ASA (5.1% vs. 11.0%; hazard ratio, 0.44; 95% CI, 0.22 to 0.87), although ASA use increased major bleeding (4.6% vs. 3.8%; hazard ratio, 1.22; 95% CI, 1.01 to 1.48). Practice guidelines recommend an approach of continuing ASA while interrupting the P2Y12 inhibitor 3 to 10 days before the planned surgery, depending on the antiplatelet drug,1,12 and recommend continuing ASA perioperatively in patients with coronary stents who require elective noncardiac surgery.1,79,95 Bridging with short-acting tirofiban, eptifibatide (glycoprotein IIb/IIIa inhibitors), or cangrelor, although not routinely recommended, may be considered in highrisk situations such as surgery within 2 to 4 weeks of coronary stenting.1 Consultation with a cardiologist and surgeon is advisable in these situations.
