A British Society for Haematology guideline on the assessment and management of bleeding risk prior to invasive procedures

Abstract

This guideline was compiled according to the British Society for Haematology (BSH) process at https://b-s-h.org.uk/. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) nomenclature was used to evaluate the levels of evidence and to assess the strength of recommendations. The GRADE criteria can be found at http://www.gradeworkinggroup.org. A literature search was carried out using the terms given in Appendix. Manuscript review was completed by the BSH Guidelines Haemostasis and Thrombosis Task Force, BSH Guidelines Executive Committee and the Haemostasis and Thrombosis sounding board of the BSH. Further review was performed by the British Society of Interventional Radiology; these organisations do not necessarily approve or endorse the contents. This guidance update from the BSH is focussed primarily on non-surgical invasive procedures, simply termed ‘procedures’ in this document, with the primary objective of giving pragmatic advice where evidence is limited. This guidance also aims to reduce unnecessary laboratory testing, inappropriate use of blood products and unnecessary delays in therapeutic procedures.1 It should be read in conjunction with the Interventional Radiology (IR) procedure bleeding risk guidance produced by the British Society of Interventional Radiology (BSIR) and the BSH.2 Recommendations are predominantly based on evidence from adult patients and therefore may not be applicable to neonates or very young children. Figure 1 gives a recommended pathway for the preprocedure assessment of bleeding risk. Published guidance ranges from stratifying procedure risk into either three tiers—low, moderate and high risk—or a more simplified dichotomy between high and low risk. Due to differences in required haematological tests between low and moderate/high risk procedures, this guideline has used a two-tier approach for bleeding risk assessment prior to elective procedures (see Figure 1). However a three tier approach remains useful for quantifying procedural bleeding risk. Comprehensive lists of relevant procedures falling into different risk categories are provided elsewhere.3, 4 It has been proposed that high-risk procedures are those with a major bleeding risk of >1.5%.5 Procedures could also be considered high risk by virtue of bleeding being more difficult to diagnose and treat (e.g. retroperitoneal bleeding versus superficial soft tissue bleeding) or with more significant consequences (e.g. bleeding secondary to spinal interventions). Procedures involving percutaneous solid organ puncture, or deep intra-abdominal drainage or biopsy, should be considered high risk. Arteriography requiring less than a 6 French sheath for access should be considered moderate risk, whereas aortoiliac or other intra-abdominal interventions such as embolisation are high risk. While the majority of venous procedures (including fistula interventions) are moderate risk, transjugular intrahepatic portosystemic shunts and thoracic venous interventions are notable exceptions and should be considered high risk. Procedural risk will also depend on the complexity of a particular procedure. On occasion, what are generally considered low-risk procedures might be more complex than usual, for example, tunnelled central venous catheter insertion in the presence of occlusive central venous disease or attempted inferior vena cava filter retrieval in the presence of filter tilt or long implantation duration. Such procedures should be considered high risk. A key recommendation of the previous BSH and the more recent European Society of Anaesthesiology (ESA) guidance is the taking of a preprocedure bleeding history.1, 6 A 2015 survey of more than 700 members of the ESA revealed that less than half of the respondents utilised a standardised history to assess bleeding risk.7 The majority of bleeding assessment tools have been developed specifically to identify subjects with an inherited bleeding disorder, most commonly von Willebrand disease (VWD).8-11 Although bleeding assessment tools have also been recommended for use in the preprocedure setting, they have not been validated and their performance remains questionable. Vries et al. found that the International Society of Thrombosis and Haemostasis (ISTH) Bleeding Assessment Tool (BAT) questionnaire in the preoperative setting did not differentiate between patients with and without defined laboratory abnormalities.12 Hence a consensus-based questionnaire, HEMSTOP (Hematoma, hEmorrhage, Menorrhagia, Surgery, Tooth extraction, Obstetrics, Parents), to assess preoperative bleeding risk was proposed.13 This tool was developed to identify adults with bleeding symptoms for whom perioperative haemostatic precautions should be considered. The HEMSTOP questionnaire contains five questions relevant to all patients and two gender-specific questions (Table 1). A HEMSTOP score of 2 or more had a specificity of 98.6% (95% confidence interval [CI], 92.3–100) and a sensitivity of 89.5% for patients requiring haemostatic precautions due to an elevated bleeding risk. With a HEMSTOP score of <2, the authors suggest that in a realistic prevalence scenario (bleeding disorder frequency of 1%), the negative predictive value would be >99%, essentially ruling out a patient-related bleeding risk requiring special precautions. The questionnaire is simple to apply and warrants further assessment and prospective validation. Although the study is too small to support a strong recommendation, it is felt that this score may have a role in the preassessment setting to indicate which patients need further haematological input. There are a number of additional caveats to consider. First, a positive score does not necessarily indicate a bleeding disorder with an estimated positive predictive value of 39%. Certain bleeding symptoms such as heavy menstrual bleeding and bruising are common in the normal population without haemostatic defects, whereas a lack of significant bleeding despite previous surgical or procedural interventions would suggest that a significant underlying bleeding disorder is unlikely. Second, there is a subjective element to the questions and so clinical judgement is required in interpretation, especially in patients already on antiplatelet agents or anticoagulants. Finally, although only 1.4% of healthy volunteers were found to have a score of 2 or more, this may not be reflective of the hospitalised population, and the capacity and resources to investigate patients identified through this tool remain unclear.14 6. Have you ever consulted a doctor or received treatment for heavy or prolonged menstrual periods (contraceptive pill, iron, etc.)? 7. Did you experience prolonged or excessive bleeding after delivery? Table 2 provides a list of antithrombotic and/or antiplatelet medications which should be highlighted in the assessment. The risk of bleeding associated with a procedure will determine the need to interrupt medication. The BSH has produced guidance for the perioperative management of anticoagulation and antiplatelet therapy.15 There is no evidence to support routine laboratory testing in patients on anticoagulants or antiplatelet agents prior to procedures other than checking the INR in patients on vitamin K antagonists. All patients on antiplatelet agents and anticoagulants should be counselled about the risks of bleeding from invasive procedures versus thrombosis associated with interruption of treatment within the consent process. One meta-analysis identified a threefold increase in major cardiac adverse events in patients who discontinued aspirin therapy given as secondary prophylaxis and a smaller study showed a similar odds ratio for ischaemic stroke.16, 17 Although these observational studies are not specific to percutaneous procedures and high-quality evidence is lacking, it is reasonable to assume some thrombotic risk to patients from pausing anticoagulants and antiplatelet agents, albeit likely <1% in most situations. Using the perioperative anticoagulant use for surgery evaluation (PAUSE) protocol in patients with atrial fibrillation requiring interruption of apixaban, dabigatran or rivaroxaban for surgery or procedures with a high bleeding risk, the overall 30-day risk of bleeding was ≤1.69% and the risk of arterial thromboembolism was ≤0.5%. Only a small minority were interventional radiology procedures and many patients will not have restarted the direct oral anticoagulant (DOAC) until 2–3 days post procedure.18 When measured, the residual anticoagulant level was <50 ng/mL in 98.8% of cases. Routine measurement of DOAC levels before procedures is therefore not indicated when using the PAUSE protocol and there is currently an absence of evidence to support a clinical utility from testing. Many low-risk procedures can be performed without pausing anticoagulants and aspirin. Procedures documented as not being associated with an increased bleeding risk if low-dose aspirin is continued include transbronchial lung biopsy, percutaneous biopsies and renal biopsy.19-22 For newer antiplatelet agents, much of the data are derived from experience in cardiac surgery rather than non-surgical procedures. In vitro experiments indicate significant differences in the duration of action and reversibility of P2Y12 inhibitors in comparison with aspirin. The effects of aspirin wear off within 4 days in comparison with 7–10 days following clopidogrel cessation.23 A greater percentage of normal platelets is required to normalise platelet aggregation in the presence of platelets inhibited by clopidogrel in comparison with aspirin. Until further clinical data become available, a conservative approach with the newer antiplatelet agents is reasonable because in many procedures, adequacy of haemostasis cannot be directly visualised and direct interventions to stop active bleeding are not feasible. After a procedure associated with immediate and complete haemostasis, for example, soft tissue biopsy without significant vascular injury, recommendations are that DOACs can be restarted at 6–8 h postprocedure. Pragmatically, this could be the patient's next routine dose beyond this time period.24 The PAUSE protocol recommends restarting the day after the procedure for low-risk procedures and a delayed restart after 2–3 days for high-risk procedures.18 No high-quality data are available to guide the timing of restarting antiplatelet agents. the timing could be as DOACs there is a significant risk of delayed can be restarted on the of the procedure or the following day at the patient's has recommended routine testing prior to elective surgical In the previous BSH et al. observational studies a positive predictive value and a ratio for that the and are of A meta-analysis predominantly of observational studies the absence of a has to the that testing is not by The clinical utility of the and as a tool is therefore This is not as the and are in vitro tests which can identify in such as in vitro inhibitors not associated with bleeding such as the and do not the complex haemostatic in patients with and In many of the bleeding are not associated with an platelet defects, and moderate of clinical has that laboratory testing in patients with a bleeding history has significant Vries et al. found that patients with and without a bleeding history prior to procedures had a similar frequency of laboratory the of many laboratory with clinical patients with a bleeding disorder remain in terms of a defined laboratory and the bleeding risk is high with such as platelet platelet haemostatic have all been used to assess bleeding risk prior to surgery or invasive procedures. This guideline does not the use of testing surgery or the use of other tests such as the time cardiac There are no prospective to small studies that with the has no predictive value for bleeding or in patients invasive procedures, those with renal Only et al. a role for in preoperative risk based on patients with The platelet the effects of aspirin inhibitors and and The of platelet following the of P2Y12 is studies have indicated that the assessment of platelet using can bleeding risk, blood This was by a review observational studies and and a meta-analysis of platelet tests for blood and in cardiac surgical there are no studies for other of The value of and in the of and the use of haemostatic support are in a BSH on in bleeding or the use of blood products in patients without disease invasive procedures, and use within the setting are limited. The level of evidence is low due to the in the of the use of different a lack of and in In a prospective study of patients in an was performed without bleeding in with normal despite increased and without the of In a study of there was a significant in blood use without bleeding after the of prior to In this if the ratio was the platelet was and the ratio was was the was normal the procedure was performed without a and platelet by and in the and and in the after the of has that and blood et al. patients patients with significant <50 to of or prior to bleeding in only post to a significant in blood use without an increase in bleeding This has been with and prior to evidence a role for in support in the management of disease with being a of the lack of validated to guide haemostatic management further to the of haemostasis in this studies have no significant from using or prior to procedures in patients with with evidence that in to In a BSH it is that the of commonly used of to or to reduce the bleeding risk, is very when the is between and has a role in a and for a haemostatic level of is an therapeutic in bleeding in the perioperative the level of required prior to an invasive procedure is to be have levels in patients with or prior to levels are generally considered to from to testing using of strength is also on The level of in haemostasis on and clinical recommendations suggest levels of at to haemostasis in a with major bleeding or an invasive there can be and in In the on the management of and in patients with it is that in vitro tests suggest an of the in patients with and that is in also suggest that should not be in the prior to elective can be with or and clinical according to their and There is no evidence to support a specific level at which should be given prior to an invasive procedure. this guideline recommends that in hospitalised should be considered if the level in a is The between platelet and bleeding risk is not and on platelet and other studies suggest that the risk of bleeding is difficult to until platelet is to There are no high-quality data quantifying the bleeding risk according to platelet in invasive procedures. In study of patients interventional procedures, platelet did not reduce bleeding or clinical In patients with platelet <50 platelet did not reduce the use of In to platelet risk of bleeding is by such as the platelet presence of and of for example, patients with were less likely to The most common for preprocedure platelet has been found to be central venous catheter There is no strong evidence to indicate that preoperative testing of platelet levels is prior to low bleeding risk procedures. with haematological that are notable platelet should be platelet is in patients with thrombotic A review of central venous catheter in patients with platelet levels of <50 no major bleeding For high bleeding risk procedures, it is reasonable to an platelet of at risk of Table for platelet for different invasive procedures. are based on or with only a small number of of small The value of platelet to levels remains especially in patients with disease and and disease is associated with in haemostatic and The is the major for the of many and anticoagulant the and are of in vitro tests of are of value in overall haemostatic in disease complex haemostatic with in and von Willebrand and of other and anticoagulant platelet of the and and have to the of in and clinical that is by clinical and laboratory to a and other such as or may on this and increase the risk of that is associated with an may also in and with platelet that bleeding In a platelet level and were the of the of major bleeding although was a key All the most invasive procedures should be these The use of INR to guide bleeding risk following invasive procedures in patients is not by clinical No clinical have for and at which invasive procedures can be considered A number of studies have the of biopsy in patients with with data that bleeding does not with of there is also evidence a between bleeding and a of INR and platelet between and the presence of venous and other to may bleeding risk following biopsy, for which the INR is a observational study of patients biopsy and haemostatic and found no between haemostatic and procedural the only identified of bleeding was 2 h post procedure. A of the literature low of bleeding in patients invasive procedures. et al. no major bleeding in patients cardiac et al. that can be performed in patients with prolonged INR secondary to either or without of et al. found that only patients significant bleeding after a of invasive procedures, biopsy, insertion and central venous although were in only of et al. a prospective study of patients with invasive were into two according to the presence or absence of as INR and/or platelet <50 No bleeding events were in either following low-risk procedures such as there was an increased risk of bleeding in the following high-risk procedures or biopsy, although this was not et al. a prospective study of invasive procedures in bleeding was and not by low platelets or prolonged The of found no in the where INR was study of showed no evidence of in of the patients the procedure (including patients with INR A study of procedures with an INR showed only 17 bleeding with no between INR and A study of found major to be with only despite INR being in A prospective observational study of procedural bleeding in hospitalised patients with or procedures an overall low of major bleeding of No was identified between platelet and procedural of bleeding were high bleeding risk procedures disease with for disease and Table 4 a list of low bleeding risk procedures and have recommended and in this haemostasis have been used prior to invasive tests to risk patients and to guide therapeutic of there is no evidence that of blood products such as or platelets haemostatic following invasive of to be at and may in patients does not after with normal despite in and may an number of the use of before invasive There is no evidence to support the use of vitamin K in patients with In a study of the majority with the in INR was following vitamin K There was no in INR between versus dose or between versus In a study of patients vitamin K for of patients with had no in the INR and those that did a were more likely to have It to be to vitamin K in patients with an increased INR secondary to vitamin K for example, or prolonged procedures should not be delayed INR if an dose has been given (e.g. and a time interval has There are no prospective studies the role of in patients and no evidence to support the use of complex or or about the with in therapy has been recommended as an in patients with bleeding when there is evidence of have not recommended routine prophylaxis prior to procedures. The study that to for h may increase the risk of venous thrombosis in patients with for a platelet which invasive procedures can be performed is limited. In vitro studies using from patients that a platelet the of the healthy In a study of biopsies in patients with bleeding was the in patients with a platelet the majority of in patients with platelet with the the platelet may also be a for risk such as and In study of patients percutaneous biopsy, the of less for preprocedure blood use and platelets was associated with than and platelets In of a of platelets in only a small increase in platelet without either or tests and may be associated with are available for use prior to elective procedures in patients with as an to platelet may be to platelet The treatment required prior to is Although not in all have been associated with an increased risk of thrombosis so should be used with in patients considered to be especially as these patients were from Although more in platelet to platelet there remains on bleeding in the active bleeding or prior to high-risk procedures of platelet and In and are In a prospective study of patients had an INR at some their are and are associated with a increased risk of for is associated with a and on is also common in this with to being at and in the A observational study in the of patients had a platelet <50 a clinical history on the may be and the tests of in haemostasis procedures the insertion of vascular access percutaneous and are common in and observational data suggest these can be carried out with a low risk of products are to patients in without evidence to support this In a observational study of of the treatment were to patients without and were to patients without bleeding and only INR Procedural prophylaxis was the documented in that bleeding in the patients is The patients with an INR of to either no or prior to central venous catheter insertion or There was no significant between the two in terms of although the study was by small The dose of was to the INR to in only of at a similar to that in with in anticoagulants as as with normal testing and normal in the can be associated with which is the of and increased to are also commonly without evidence to support this The study of patients found received platelet their the treatment of patients had a platelet at the time of and of patients were not bleeding at the time of The in platelet was were in a observational in with of platelet given to simply procedural There is a lack of evidence to support platelet in the A recent of platelet versus no prior to central venous catheter insertion in patients with a platelet of (including a in major bleeding in the platelet as suggest platelet bleeding specifically in patients either insertion or in those with platelet bleeding were in patients tunnelled A small in patients with no in blood between those platelet and those in has been associated with adverse increased The risk of bleeding may also be increased or by and as as and of haemostatic and medication. to reduce the risk of procedural bleeding should be considered (Table A meta-analysis access with and without the use of guidance showed a significant in vascular when guidance was with number of and have also been in central venous catheter insertion when using with the Although some in arterial or venous there are no data to suggest a in terms of bleeding risk This is also by et al. percutaneous intra-abdominal interventions and et al. percutaneous showed no significant increase in major bleeding when were It has been that procedures requiring an arterial access sheath of more than French should be considered high risk. there are data vascular are access evidence the use of guidance is associated with a bleeding risk for vascular organ biopsy and of frequency overall for venous access not for bleeding risk for and venous access approach to risk approach can be considered for biopsy in patients with increased risk of bleeding risk of biopsy surgical procedures should not be delayed with history of oral of a risk increased risk of bleeding biopsy and between renal disease and experience and with particular are also have a of access bleeding when this approach is used by who primarily use access for of major bleeding are also documented at et al. considered procedures to be the for of central venous catheter with a meta-analysis of bleeding post biopsy highlighted data on with study showed a bleeding from to be to their less this is to be by a more risk of the All authors to and the the BSH Haemostasis and Thrombosis Task members the time of this guideline were and The authors would to the BSH sounding and the BSH Guidelines Committee for their support in this The BSH the the of this All authors have a of to the BSH and Task which may be on The authors have no relevant of to to this of the will the if evidence available that would the strength of the recommendations in this or it The will be by the relevant Task The will be and from the BSH if it recommendations are an will be on the BSH While the advice and in this guidance is to be and at the time of to the the BSH the for the of this are to to of used terms also and search surgical invasive interventional percutaneous blood blood bleeding bleeding risk bleeding bleeding questionnaire surgical blood blood vitamin prevalence vitamin vitamin K