Peri‐operative management of anticoagulation and antiplatelet therapy

Abstract

The Grading of Recommendations Assessment, Development and Evaluation (GRADE) nomenclature was used to evaluate levels of evidence and to assess the strength of recommendations. The GRADE criteria are specified in the British Committee for Standards in Haematology (BCSH) guidance pack (http://www.bcshguidances.com/BCSH_PROCESS/42_EVIDENCE_LEVELS_AND_GRADES_OF_RECOMMENDATION.html) and the GRADE working group website http://www.gradeworkinggroup.org. Details of the literature review are given in Appendix 1. Review of the manuscript was performed by the British Society for Haematology (BSH) Guidelines Haemostasis and Thrombosis Task Force, BSH Guidelines Executive Committee and by the Haemostasis and Thrombosis sounding board of the BSH. The latter comprises 50 or more members of the BSH who have commented on the content and applicability in the UK setting. It has also been sent to the following organisations for review: The Royal College of Surgeons; The Royal College of Anaesthetists; Thrombosis UK, a patient-centred charity dedicated to promoting awareness, research & care of thrombosis; the British Dental Association; IntraHealth (who operate National Health Service General Practitioner and community pharmacies); and the British Cardiovascular Society; these organisations do not necessarily approve or endorse the contents. A BSH guideline on warfarin (Keeling et al, 2011) addressed the issue of perioperative management and is updated in this article to include the issue of perioperative management of patients on direct oral anticoagulants (DOACs) and antiplatelet agents, which are becoming frequent clinical queries. This guideline will consider whether and when anticoagulants and antiplatelet agents should be stopped before elective surgery and invasive procedures, when agents can be restarted and how to manage patients on these drugs who require emergency surgery. If an anticoagulant or antiplatelet effect persists, haemostasis may be improved by the use of pre-operative parenteral tranexamic acid, which has been shown to reduce blood loss and transfusion requirements in both cardiac and trauma surgery, without increasing thrombotic complications (McIlroy et al, 2009; Shakur et al, 2010). For agents with a slow offset and onset of action, bridging therapy with an alternative drug at a full treatment dose can be considered in patients deemed to be at high risk of thrombosis; this mainly concerns whether treatment dose low molecular weight heparin (LMWH) or unfractionated heparin (UFH) should be given when warfarin is temporarily discontinued. Thromboprophylaxis with low dose LMWH is not regarded as ‘bridging’. For some invasive procedures, such as dentistry (Perry et al, 2007) (see also http://www.sdcep.org.uk/published-guidance/anticoagulants-and-antiplatelets/), joint injections (Ahmed & Gertner, 2011), cataracts (Jamula et al, 2009), pacemaker insertion (Ahmed et al, 2010; Airaksinen et al, 2013) and certain endoscopic procedures (Veitch et al, 2016), anticoagulation may not need to be stopped. Procedures that require anticoagulation to be stopped will vary in their bleeding risk and, importantly, the consequences of bleeding will depend on the site of surgery and local anatomy. Although some have grouped procedures into lower or higher risk (Spyropoulos & Douketis, 2012; Baron et al, 2013) we think the operating surgeon, dentist, or interventional radiologist has to assess the risk of bleeding for the individual patient and discuss both this and the plan for peri-operative anticoagulation with them. The plan must be recorded clearly in the notes, including a plan for when the patient is discharged. Warfarin has a half-life of approximately 36 h and as its effect subsides γ-carboxylated vitamin K-dependent procoagulant factors need to be synthesised. Warfarin therefore needs to be stopped 5 days before elective surgery to ensure haemostasis has returned to normal. This is likely to differ for other vitamin K antagonists with different half-lives (acenocoumarol, 10 h; phenindione, 8 h; fluindione, 3 days; phenprocoumon, 5 days). If possible, the International Normalized Ratio (INR) should be determined the day before surgery to allow the administration of phytomenadione if the INR is ≥1·5, so reducing the risk of cancellation. The INR should be checked on the day of surgery. Stopping warfarin for a shorter time and attempting to reverse its effect with oral phytomenadione on the day before surgery did not prove a satisfactory alternative (Steib et al, 2010). Due to its slow onset of action warfarin can be resumed, at the normal maintenance dose (Douketis et al, 2012), or with two initial days of double maintenance dose (Schulman et al, 2014), the evening of surgery (or the next day) if there is adequate haemostasis. There have been many reviews and attempts to estimate the risk of peri-operative thrombosis (Dunn & Turpie, 2003; Dunn et al, 2007; Dentali et al, 2012; Douketis et al, 2012; Siegal et al, 2012; Spyropoulos & Douketis, 2012). The main question has been whether the risk of thrombosis is sufficiently high, in patients who have temporarily discontinued a coumarin, to use treatment dose LMWH or UFH pre-operatively and/or post-operatively when haemostasis is secure. This is predicated on the assumption that bridging will reduce the thrombotic risk. It is noteworthy that in their meta-analysis Siegal et al (2012) found no difference in the risk of thromboembolic events in eight studies comparing bridged and non-bridged groups of patients (odds ratio [OR], 0·80; 95% confidence interval [CI], 0·42–1·54). However, bridging was associated with an increased risk of major bleeding in five studies (OR, 3·60; 95% CI, 1·52–8·50). A further systematic review also concluded ‘while the antithrombotic efficacy of perioperative bridging with LMWH has not been demonstrated, increased bleeding risk is observed in different types of surgery’ (Eijgenraam et al, 2013). Patients taking warfarin for the treatment and secondary prevention of venous thromboembolism (VTE), for stroke prevention in atrial fibrillation (AF) or for mechanical heart valves (MHV) need separate consideration. For patients with acute VTE the risk of recurrence without anticoagulation is very high in the first 3 months (Kearon & Hirsh, 1997) and surgery will increase the risk further. When a patient is more than 3 months from an acute event the risk of recurrence is much lower, the treatment phase is over, and patients are remaining on an anticoagulant for secondary prevention (Kearon & Akl, 2014). Prophylactic dose LMWH can substitute for warfarin after the acute treatment period, hence patients with VTE more than 3 months prior can usually simply be given post-operative prophylactic dose LMWH (or a suitable alternative) rather than receive full dose bridging therapy while anticoagulation with a coumarin is re-established. Bridging might be considered for those thought to be at very high risk, such as patients with a previous VTE occurring whilst on therapeutic anticoagulation who now have a target INR of 3·5. For patients with a MHV the risk varies with type of valve (bileaflet less than caged ball and tilting disc), valve position (aortic less than mitral) and patient risk factors (such as previous stroke or transient ischaemic attack (TIA), AF and reduced left ventricular ejection fraction). We have previously recommended bridging therapy for patients with MHVs other than those with a bileaflet aortic valve and no other risk factors (Keeling et al, 2011) and there is no strong evidence to change this recommendation. For patients with atrial fibrillation the CHADS2 (Congestive failure; Hypertension; Age ≥75 years; Diabetes mellitus; prior Stroke, TIA or thromboembolism) score or, more recently, the CHA2DS2VASc CHADS2 + Vascular disease; Age 65–74 years; Sex category) score has been used to predict stroke risk. The CHADS2 score may also predict risk of post operative stroke (Kaatz et al, 2011) and guidelines have suggested the CHADS2 score is used to select patients for bridging (Douketis et al, 2012) whilst the previous BCSH guideline suggested bridging in only those with a previous stroke or TIA or multiple other risk factors (Keeling et al, 2011). There is now a randomized, double-blind, placebo-controlled trial of bridging in AF patients (Douketis et al, 2015). Patients were randomised to dalteparin (100 iu/kg bd) or placebo from 3 days before until 24 h before the procedure and then for 5–10 days after the procedure. A total of 1884 patients were enrolled, with 950 assigned to receive no bridging therapy and 934 assigned to receive bridging therapy. The incidence of arterial thromboembolism was 0·4% in the no-bridging group and 0·3% in the bridging group (risk difference, 0·1 percentage points; 95% CI −0·6 to 0·8; P = 0·01 for non-inferiority). The incidence of major bleeding was 1·3% in the no-bridging group and 3·2% in the bridging group (relative risk, 0·41; 95% CI 0·20–0·78; P = 0·005 for superiority). The authors concluded forgoing bridging was non-inferior to bridging for the prevention of arterial thromboembolism and decreased the risk of major bleeding. Thirty-eight per cent of patients had a CHADS2 score of ≥3, though only 3·1% had a score ≥5 (which requires a previous stroke or TIA plus at least three of four other risk factors), 9·4% had a previous stroke and 8·3% a previous TIA. Patients with a stroke or TIA within the previous 12 weeks were excluded. Our updated advice is shown in Table 1. When we say ‘consider bridging’ we do not mean it should be automatically given, but consider whether to bridge or not in discussion with the patient. In patients who are receiving pre-operative bridging with LMWH, the last dose should be at least 24 h before surgery and if on a once a day regimen, some recommend the last dose is halved for high risk surgery (Douketis et al, 2012). We recommend that post-operative bridging (i.e. full dose anticoagulation) is not started until at least 48 h after high bleeding risk surgery although thromboprophylaxis should be given if indicated. If surgery can wait for 6–8 h then 5 mg of intravenous phytomenadione can restore coagulation factors; if this is not possible, anticoagulation can be reversed with 25–50 u/kg of four-factor prothrombin complex concentrate (Refaai et al, 2013; Goldstein et al, 2015), we would give at the lower end of this range and check the INR. Post-operative management should follow the same strategy as for elective surgery. The approach to the peri-operative management of patients on Direct oral anticoagulants (DOACs) is based on an approximate calculation of the half-life of the drug and therefore its persistence in the circulation, taking into account renal function. This is combined with consideration of the bleeding risk of the proposed procedure and a clinical evaluation of the patient's individual risk factors for thrombosis and bleeding. Current strategies for elective surgery do not routinely include measurement of either non-specific or specific coagulation parameters to assist in quantification of DOAC levels. For each of the drugs there are data on periods of discontinuation during the studies to evaluate the drug against coumarin therapy. In addition there is a single moderate sized, prospective study evaluating the outcomes of a set for the management of patients on (Schulman et al, 2015). the Evaluation of study were discontinued in et al, 2012), of the study A range of procedures was performed cardiac and but also more major surgery. The incidence of major bleeding was and for mg and warfarin and the of stroke and was in each group for the the bleeding and thrombosis were more after emergency and major A prospective study that risk and high risk procedures and which the in Table low of major and in the the procedure (Schulman et al, 2015). was restarted post-operatively only when haemostasis had been procedures it at mg on the evening of the procedure to or mg the following For major procedures the was by with patients at normal full dose h Bridging with LMWH or UFH was used in of but not at In this of the procedures had a high bleeding risk and eight of the 10 patients who had major bleeding had a high bleeding risk procedure. from the study that in patients procedures, periods of discontinuation of for days to allow surgery or an invasive procedure in no difference in major stroke and and when to discontinuation of warfarin et al, 2014). There are no data that on a for discontinuation of but groups have proposed for this et al, et al, 2015). The advice is to for 24 and 48 h for low bleeding risk and high bleeding risk from the trial also procedures anticoagulation may be et al, 2014). The of the events were low risk procedures, and was discontinued for of the major outcomes were in patients taking warfarin and of whether or not anticoagulation was discontinued. in this when was discontinued it was for There are no data that on a for discontinuation of The advice is to for 24 and 48 h for low bleeding risk and high bleeding risk on the discontinuation of give different et al and et al renal into consideration and periods of discontinuation for those with of high risk procedures of 48 h while et al did not consider renal but recommend discontinuation for high bleeding risk procedures for at least In we have considered the of patients periods of et al and the guideline et al, a of discontinuation for 36 h in patients with of low risk procedures, this is not for patients taking a once such as and so we have to recommend a of discontinuation for this group when either or The oral to be in the UK is of which is and its half-life is The of that for or other procedures it should be discontinued at least 24 h before the procedure. There are data on the management of emergency surgery in patients receiving The to haemostasis at surgery in these patients is by in the of each drug that is associated with a UK National Service in found high of for the and for the when levels of UK The was in the measurement of low of and, drug was as in there was no drug If an anticoagulant effect be should be When possible, surgery should be to allow the of the drug to The of drug can be from the dose of the time of last dose and the renal function. such as patient weight and the use of will also have less The approximate half-life of the drugs to renal is given in Table of coagulation such as the prothrombin time time and time may allow an approximate estimate of the levels of drug in the et al, 2014). A normal the of in a but normal and do not the of of or in a would then that or can be to the but with given to the of the in of the et al, 2014). It has been suggested that use of agents might reduce the risk of bleeding in patients on who require emergency surgery. of the evidence in this to and to the of in in of haemostasis. The of is that treatment with a prothrombin complex concentrate might but this into account the for thrombotic which are There are data that the use of and in the management of emergency surgery and so a approach might be to with surgery only in the event of bleeding. is likely to reduce bleeding and should be management strategies include further of anticoagulants and the use of such as drugs and and that might further haemostasis. is and so can be by if a procedure can be for for this to but this is of can be by a single although in have been observed on of et al, 2013; et al, 2015). This strategy is not to and which are In a prospective study of the administration of a dose of 5 of to 36 patients who were receiving prior to an invasive or normal haemostasis was in and or haemostasis was in patients and thrombotic event within h after In the same major of effect on coagulation was observed in patients given in of patients et al, 2015). There are no data on the use of in patients surgery but there are now data on the of anticoagulation in In a study of and who a dose of in anticoagulant were in both groups et al, 2015). In was reduced by with those who placebo and by with was in of and of placebo In was reduced by with those who and by with was in of and of placebo were when was as a plus an or thrombotic events were Although these data be as to haemostasis during surgery, the are et al, 2015). therapy is a in the secondary prevention of This to following ischaemic and antiplatelet therapy following acute when a of and an is after In these the increase in bleeding risk associated with and more so with et al, 2009), is by their clinical However, the of antiplatelet agents in the is associated with an increase in bleeding risk. In a meta-analysis of studies et al that therapy was associated with a increase in post-operative bleeding but no increase in the of that low dose be procedures and A meta-analysis has shown that patients on who have a can be by normal with surgery et al, However, the same may not be for which is associated with more bleeding to et al, 2012). This bleeding risk has to be against the increased thrombotic risk associated with or of et al, or et al, 2013; et al, 2015), which may be to a or other invasive procedure. in elective surgery have peri-operative or of in patients with very had or had a procedure et al, 2010; et al, et al, 2014). of the studies were with and were therefore to assess in bleeding events et al, 2010; et al, 2011). from 10 days prior to surgery until the of surgery in no increase in thrombotic events or in bleeding events et al, 2011). In from days prior to surgery until 3 days in a higher of P = but no difference in peri-operative blood loss et al, 2010). in guidelines have recommend of surgery is to have a high bleeding risk or is in a such as or et al, 2011). a including patients on an antiplatelet no difference in the of or ratio in major bleeding if was as to from day until days et al, 2014). bleeding was increased in a separate group not on but to it There is very and less data on bleeding and of single et al, some guidelines in the same as with et al, 2012), we do not there is evidence to a recommendation. is the management of invasive This is an as more patients with insertion following when is recommended for at least weeks following and 12 months following shorter is with the and of patients will require surgery within of The risk of peri-operative is within the first after with risk at months and et al, et al, 2011). for include of antiplatelet or high surgery and et al, et al, 2015). There are no on which to advice in this hence guidelines a a the patient's thrombotic risk and the bleeding risk associated with the type of invasive procedure et al, et al, 2014). In very low procedures can be without low risk procedures in patients with low thrombotic risk may be on with of the elective surgery in patients deemed to be at high thrombotic risk should be until are lower risk. If surgery be then it should on with discontinuation of the In high thrombotic risk patients high surgery that be consideration can be given to bridging with a parenteral such as or during the of et al, 2011). The of following surgery is In a meta-analysis of studies the use of was associated with a in ischaemic events but with in bleeding and transfusion requirements and a increased risk of surgery et al, 2015). In in a the administration of to patients surgery did not in a lower risk of or thrombotic complications or in a higher risk of bleeding as to placebo et al, was also in a meta-analysis of to increase and the for blood as as to increase in et al, In both these the study groups were and in the it is that many of the were also on many of the studies were when use of tranexamic acid, which has been shown to reduce blood loss in both and patients (McIlroy et al, 2009; et al, was less cardiac guidelines recommend there is a very high bleeding risk, a very low thrombosis risk or the patient would transfusion et al, 2012; et al, 2014). surgery combined and is by a further increase in to to and and had bleeding while and was associated with higher bleeding and et al, et al, 2012). based on data on bleeding and drug it is recommended that and are discontinued 5 days and days while is et al, 2012; & 2013; et al, 2014). In high thrombotic risk bridging with a parenteral antiplatelet during of the oral have been proposed et al, & 2013). The parenteral is usually on day and stopped h and h until the can be restarted when bleeding has been and in the of an initial dose is recommended et al, 2014). When high surgery is and time not of or both antiplatelet agents, there is evidence from both in et al, 2007; et al, 2012) and in et al, 2012) studies that transfusion of may haemostasis. should be at least h after the last dose of and h after the last dose of to by drug or of requires and is more can be by in so whilst of reverse the effect of the effect of higher of when on antagonists is less certain et al, 2007; et al, 2012; et al, et al, 2015). A single dose of in patients with an on did not et al, in to may allow some patients to have a shorter of discontinuation pre-operatively 3 days for or and 5 days for This is in surgery when of may those patients in which surgery may be et al have the of A of are the clinical evidence to be with or multiple the a normal to was associated with low bleeding et al, 2011). of was associated with an higher risk of transfusion while a had a of for at least transfusion et al, 2010; et al, 2014). et al on of the Society of studies of in a of and Although has been following or such events are incidence in and in In a of of antiplatelet therapy was only in three and no of were in combined of patients while on antiplatelet therapy. the guideline and can be without of or et al, 2010). In a review of this was for but not for management procedures et al, 2015). evidence is in to the of in patients receiving and therefore it is recommended that such agents be discontinued days prior to the procedure et al, 2010; et al, 2015). the advice and in this guidance is to be and at the time of to the the the for the content of this the authors the literature and the initial of the The authors to for in the initial literature The BSH Haemostasis and Thrombosis members at the time of this guidance were A K and The authors would to and the BSH sounding BSH Guidelines for their in this The BSH Guidelines the during the of this guidance (see has been for by and has been for by and of the group will the group if evidence that would the strength of the in this or it The will be and from the BSH guidelines website if it If are an will be on the BSH guidelines for to clinical question only oral anticoagulants only K antagonists only and antiplatelet and anticoagulant based on only of to or