Drugs to reduce bleeding and transfusion in major open vascular or endovascular surgery: a systematic review and network meta‐analysis

Objectives

This is a protocol for a Cochrane Review (intervention). The objectives are as follows:

To assess the efficacy and safety of anti‐fibrinolytic and haemostatic drugs and agents in reducing bleeding and transfusion in people undergoing major vascular surgery or vascular procedures with a risk of moderate or severe (> 500 mL) blood loss.

Background

Description of the condition

Vascular surgery treats diseases of arteries, veins or lymph vessels, except for those in the heart or brain. The major types of arterial disease are arterial aneurysms, arterial dissections and arterial occlusive disease. These conditions may be treated either with open surgery by vascular surgeons, or with endovascular procedures conducted by vascular surgeons or interventional radiologists (Hirsch 2006). The availability of interventional radiology varies regionally and globally, depending on the availability of trained staff and equipment (Beck 2016; Kline 2017).

Aneurysms

Aneurysms are abnormal dilations in an artery which can progressively enlarge and weaken, with risk of rupture and severe internal bleeding. Vascular services manage aneurysms found in the chest (thoracic aortic aneurysm, TAA), chest and abdomen (thoraco‐abdominal aortic aneurysm, TAAA) or abdomen (abdominal aortic aneurysm, AAA) as well as aneurysms found in peripheral arteries, including iliac, popliteal or femoral arteries. Aneurysms can remain asymptomatic, but rupture can be fatal or life‐threatening. Aneurysm repair can be conducted as an emergency in the case of leak or rupture, or electively to prevent rupture and other complications. Elective repair of AAA is associated with significantly reduced mortality of 2% compared to emergency repair mortality of 20% to 30% (Majd 2016). AAA is the most common type of aneurysm to require repair, whether with open surgery or endovascular repair (Sampson 2014). Risk factors for AAA include older age, male gender, European ancestry, smoking and high blood pressure (Altobelli 2018). Ultrasound screening programmes estimate AAA prevalence at 2.7% in 65 to 74 year olds and 7.3% in 75 to 85 year olds (Makrygiannis 2016). The Centers for Disease Control and Prevention (CDC) ranks AAA as one of the top 15 causes of mortality in the USA for those between 85 and 89 years (CDC 2015).

Arterial dissection

Arterial dissection is a process in which blood tracks between the layers of an artery wall, forcing them apart. This can be an acute or chronic process, initiated by a defect in the vessel due to shear stress, inflammation, trauma or at the site of an aneurysm. Seventy per cent of major vessel dissections occur in the ascending thoracic aorta (Stanford type A), 7% in the thoracic arch, 20% in the descending thoracic or thoraco‐abdominal aorta (Stanford type B) and 2% in the abdominal aorta (Howard 2014; Roberts 1991). Stanford type A requires urgent or emergency open surgical or endovascular repair, whilst Stanford type B may also be managed by reducing blood pressure with medications (Bannazadeh 2016; Cooper 2016; Elsayed 2017; Ulug 2012). Risk factors for aortic dissection include inherited or acquired connective tissue disorders, high blood pressure and aortic aneurysm (Paterick 2013). Acute aortic dissection has an estimated incidence of 52 per 100,000 per year, 60% occur in men, and it has a high risk of death (approximately 73% 30‐day mortality if Stanford type A, and 13% mortality if Stanford Type B) (Howard 2013). If the aortic valve is involved in thoracic aortic aneurysm or dissection, surgical repair may require cardiac surgeons.

Occlusive arterial disease

Occlusive arterial disease (OAD) is caused by atherosclerosis, in which fat and cholesterol deposition cause inflammation, thickening and hardening of vessel walls, with eventual narrowing or blockage of the artery (Rahman 2017). This can cause inadequate blood flow through the vessel, with poor oxygen delivery to tissues beyond it (ischaemia). Increased blood flow through other (collateral) vessels may compensate to some degree (McDermott 2017). Atherosclerotic deposits (plaques) can also rupture, suddenly blocking blood flow with blood clot (thrombus) and debris, causing sudden and severe (critical) ischaemia and resultant tissue, organ or limb death (Gilliland 2017). OAD risk factors include being male, older age, personal or family history, cardiovascular disease, diabetes, stroke, high cholesterol, high blood pressure, smoking, obesity and inactivity. OAD can occur in many locations, but vascular services typically treat blockage or narrowing (stenosis) of arteries in the neck (carotid arteries), abdomen (aorta), pelvis and legs (iliac, femoral or popliteal arteries). Carotid stenoses or occlusions cause 15% to 25% of strokes (Saw 2014). Carotid OAD can be managed with open surgery (endarterectomy) or endovascular stents (Noiphithak 2017). Occlusive disease in the aorta is classified as above (supra‐renal) or below (infra‐renal) the artery to the kidneys and can also be managed with open bypass grafting surgery or endovascular stenting.

OAD most commonly affects the lower limbs, where it is also known as peripheral arterial disease (PAD) (Fowkes 2017). It is defined by ankle to brachial (upper arm) blood pressure index (ABPI) of less than 0.9. The prevalence of asymptomatic PAD in the middle‐aged to elderly population is estimated at 7% to 15%, and affects over eight million Americans (Swaminathan 2014). The PERART study (a Spanish primary care population) found a PAD prevalence of 10.2% in males and 5.3% in females (Alzamora 2010). The National Health and Nutrition Examination Survey (NHANES, 1999 to 2000) reported a symptomatic PAD prevalence of 4.3% in adults aged over 40 years old and 14.5% in adults over 70 years old (Selvin 2004). However, the British Regional Heart Study, using femoral artery ultrasound assessment, found 64% of subjects aged 56 to 77 years had significant femoral atherosclerosis, of which only 10% were symptomatic (Leng 2000). When vascular disease causes ischaemia with resultant tissue death, vascular surgeons may perform an amputation at the lowest unaffected level, for example below‐knee, above‐knee or hind‐quarter amputation. Retrospective studies show that non‐traumatic amputations are nearly all caused by vascular disease, which may or may not be complicated by diabetes, and have a high risk of death (mortality of 30% at 30 days and 54% at one year) (Kristensen 2012).

Open vascular and endovascular procedures

Open aneurysm or bypass surgery, particularly in the chest, abdomen and pelvis, are invasive major operations associated with complications including bleeding, stroke, cardiac and kidney injury and spinal cord ischaemia, with relatively long recovery and length of hospital stay, and high readmission rates (Fry 2018; Hobson 2018). They may also require periods of aortic cross‐clamping, which adds to complication rates (Zammert 2016). Where feasible, procedures are conducted endovascularly with stents and grafts, guided by contrast dye and radiological imaging. Endovascular procedures avoid large incisions, cause less postoperative pain, and may have lower mortality and complication rates with reduced hospital length of stay and costs. Intra or postoperative bleeding can occur in endovascular surgery from the vascular access site, around the graft (endoleak), or from vessel rupture. Conversion to open surgery or repeat endovascular procedures are sometimes necessary. Endoleaks (Types I to V) are defined as a persistent blood flow outside the lumen of an endoluminal graft but within the aneurysm sac or the adjacent vascular segments (Society for Vascular Surgery). They are caused by incomplete sealing or exclusion of the aneurysm sac. Endovascular procedures are now feasible in most elective and some emergency settings, particularly in high‐income countries, but remain inappropriate for some complex procedures and require expertise and equipment that may not be available in some settings (Buck 2014).

A 2014 Cochrane Review found that for elective AAA repair, endovascular aneurysm repair (EVAR) was associated with lower short‐term mortality than open surgical repair (OSR), particularly with regard to respiratory complications. At intermediate and long‐term follow‐up, however, they performed comparably. Additionally, individuals undergoing EVAR had a higher re‐intervention rate to manage endoleaks, but these were mostly catheter‐based interventions associated with low mortality and were not associated with any difference in terms of 30‐day mortality (Paravastu 2014). Elective AAA repair may also be conducted with laparoscopic (keyhole) surgery (Robertson 2017). A 2017 Cochrane Review found that for urgent or emergency repair of ruptured AAA, EVAR and OSR had similar 30‐day mortality rates, but did not find a difference in complication rates (Badger 2017). A 2016 Cochrane Review found no randomised controlled evidence to support thoracic endovascular aneurysm repair (TEVAR) compared to open surgical thoracic aorta repair; observational studies, however, support the use of TEVAR (Abraha 2016). A 2017 Cochrane Review found that endovascular treatment (percutaneous transluminal angioplasty) for chronic limb ischaemia was associated with fewer early complications and shorter hospital stay compared with bypass grafting surgery. However open surgical treatment had better flow in vessels one year on. Endovascular treatment of lower limb occlusive arterial disease may therefore be particularly beneficial in people with significant comorbidities which make them high‐risk surgical candidates (Antoniou 2017). In general, patient selection ‐ surgical or interventional radiologist experience and anatomy of the defect ‐ determine whether endovascular or open surgery is preferable.

Bleeding and transfusion in vascular procedures

Internal bleeding may occur before surgery (in the case of arterial dissection or rupture), during an intervention, or after an intervention, due to inadequate surgical haemostasis, abnormal clotting, graft failure, migration or endoleaks. Bleeding ranges from minor, with no transfusion requirement, to massive, requiring multiple blood product transfusions. Operations or procedures with a risk of at least 500 mL blood loss include open or endovascular emergency repair of AAA, TAAA or TAA, open or endovascular repair of thoracic aortic dissection, complex lower limb bypass surgery, and major lower limb amputation. Studies show transfusion rates of 38% in people undergoing elective open AAA repair, 27% in lower limb bypass surgery, 15% in open thromboendarterectomy (removal of blood clot and atherosclerotic plaque), and vary from 17% to 64% in lower limb amputation (D'Ayala 2010; O'Keeffe 2010; Tan 2015). Some procedures, for example endovascular AAA repair or endovascular lower limb stenting, have a low risk of bleeding but are performed frequently, and therefore account for more blood product transfusion overall (Obi 2015). Other procedures very rarely experience bleeding or require transfusion, for example a transfusion rate of less than 1% in elective carotid endarterectomy (Rubinstein 2013). Importantly, transfusion rates and transfusion practices vary between centres and care providers (Osborne 2018).

Various surgical factors can increase risk of bleeding, including emergency procedures, for example for aneurysm rupture, dissection or critical limb ischaemia, revision or repeat surgery or complex or branching anatomy (Obi 2015). Perioperative factors include systemic anticoagulation with heparin to prevent graft thrombosis or clot extension, pre‐existing use of anticoagulants or antiplatelet drugs, intraoperative hypothermia, cross‐clamp position, acute coagulopathy in the setting of trauma, and systemic inflammatory response in the setting of an infectious disease, for example aortic mycotic abscess (Obi 2015; Samoila 2017). Procedure‐specific models have been developed to predict bleeding for certain vascular procedures (Kapma 2017; Mahmood 2018). For example, using data from a large multi‐centre quality improvement database, transfusion rate within 24 hours of EVAR was predicted at 3.2%, with the following risk factors associated with transfusion: haematocrit less than 36%, increased aortic diameter, functional status and chronic obstructive pulmonary disease (O'Donnell 2018).

Interventions to reduce bleeding and allogenic transfusion

Cell salvage can be used to collect blood from the surgical field for autologous transfusion, and meta‐analysis of vascular surgery randomised controlled trials suggest cell salvage reduces perioperative transfusions by up to 37% (Ashworth 2010; Takagi 2009). However 30% to 50% of surgical blood loss is absorbed into swabs, therefore swab washing can increase blood available for autotransfusion (Haynes 2005). Using a range of techniques to reduce bleeding may be more useful than cell salvage alone; moreover cell salvage is not used during endovascular procedures. The ratio of different blood products transfused also appears to be important to people's outcomes, as well the overall amount of blood product transfused (using a higher or lower transfusion trigger) (Mesar 2017). Strategies to reduce use of any allogenic blood products include techniques such as arterial cross‐clamping, medications to reduce blood pressure and thus reduce bleeding, and limiting use of crystalloid or other fluid infusions, which can compound bleeding by diluting clotting factors present in the circulation (Chee 2016). In addition, point‐of‐care viscoelastic testing (rotational thromboelastometry (ROTEM) or thromboelastography (TEG)) quantifies coagulation and fibrinolysis parameters and their use can guide and reduce autologous transfusion, though most evidence is in the context of cardiac surgery (Wikkelsø 2016). Finally various haemostatic drugs, which alter coagulation and fibrinolysis, are an important part of management to reduce bleeding and transfusion risk.

Description of the intervention

When an injury occurs, the formation of a blood clot (normal haemostasis) stops excessive bleeding. Blood clot formation is initiated by tissue injury, endothelial and collagen exposure and release of factors which cause blood vessel constriction (vasoconstriction) and platelet activation (Blanco 2017). Activated platelets stick together, forming a weak plug (Mackman 2007). Multiple enzyme pathways are also activated and amplified, finally producing thrombin, an enzyme that converts fibrinogen to fibrin. Fibrin rapidly polymerises and crosslinks with platelets to form an insoluble, stable blood clot. The clot is further stabilised and contracted by cross‐linking between the fibrin strands by factor XIII (Chapin 2015).

To prevent harmful, unregulated clot extension beyond the injury, blood clots are subsequently contained and broken down by fibrinolysis (Blanco 2017). The enzyme plasmin, a protease, cuts through the fibrin mesh, releasing soluble fragments that are metabolised in the liver and kidneys (Hudson 2017). Plasmin is activated locally from its precursor, plasminogen, as part of the normal clotting process. Plasmin formation and fibrinolytic processes normally occur more slowly than coagulation, such that clot breakdown occurs well after clot formation and tissue remodelling—that is, after bleeding has stopped (Chapin 2015). To prevent plasmin digesting non‐clot tissue or proteins, plasminogen is predominantly converted to plasmin at the site of and within the blood clot, creating bound, rather than free plasmin. Free plasmin will indiscriminately digest plasma proteins, including clotting factors and is normally kept in check and neutralised by circulating alpha‐2‐plasmin inhibitor (Madurska 2018). This reduces pathological, rather than physiological fibrinolysis (Makar 2010).

Antifibrinolytic drugs inhibit the activity of plasmin and thus reduce the breakdown of fibrin within blood clots, resulting in greater early and persistent clot strength (Okamoto 1997). Haemostatic drugs are a broad class of drugs which each act on distinct parts of the coagulation cascade to replace or enhance missing or poorly functioning pro‐coagulant enzymes, substrates or factors. These could be deficient due to inherited conditions, such as haemophilia, or acquired conditions, such as prolonged bleeding (consumption of clotting factors), liver failure, autoimmune disease or drug therapy.

Antifibrinolytic drugs

Tranexamic acid (TXA)

TXA is a synthetic analogue of the amino acid lysine. It binds reversibly to lysine receptor sites on plasminogen, prevents activation of plasminogen into plasmin, and reduces fibrin breakdown. This improves clot formation, stability and duration. TXA has been well validated for use in perioperative, obstetric and trauma care, as well as in cardiac surgery (Henry 2011; Ker 2015; Shakur 2018). A systematic review and network meta‐analysis of antifibrinolytic adverse drugs effects in the setting of cardiac surgery suggests TXA use reduces mortality compared to placebo or aprotinin. In addition, it does not increase myocardial infarction (MI), cerebrovascular attack (CVA) or renal failure or dysfunction (Hutton 2012). In high doses, however, TXA has been associated with seizures in the cardiac surgery setting (Murkin 2010).

ε‐aminocaproic acid (EACA)

EACA is another synthetic lysine analogue, with a similar mechanism of action to tranexamic acid. Comparative potency of EACA and TXA estimates vary but suggest EACA is 7 to 10 times less potent than tranexamic acid (Thomsen 2006). There is no known association with seizures.

Antifibrinolytic drugs such as EACA and TXA are usually administered intravenously after induction of anaesthesia. Usually a loading dose is given followed by continuous infusion. High doses appear to be more effective than low doses (Henry 2011). Neither TXA nor EACA has been associated with increased risks of adverse effects (Hutton 2012).

Aprotinin

Aprotinin is an enzyme inhibitor with complex effects on haemostasis. It is a competitive inhibitor of various serine proteases, including plasmin and kallikrein (McCarthy 1994). Plasmin inhibition slows the rate of fibrinolysis. Aprotinin exerts a much greater effect on free plasmin, however, with much less effect on bound plasmin. This improves the haemostatic problems caused by excessive or unregulated free plasmin activity, such as consumption of clotting factors. This reduces pathological rather than physiological fibrinolysis (Royston 2015). Kallikrein inhibition reduces factor XIIa activity, which inhibits intrinsic coagulation pathways leading to the formation of thrombin and fibrin. On balance, aprotinin is frequently classed as antifibrinolytic, as it has a net clot‐stabilising effect which outweighs its kallikrein‐mediated anticoagulant effects.

Aprotinin has been associated with a higher rate of adverse effects than the lysine analogues (Henry 2009). Evidence from three observational studies and from a single randomised study, in adults undergoing cardiac surgery, showed an increased risk of renal dysfunction, cardiovascular events, pulmonary embolism and death with aprotinin (Bremerich 2006; Cooper 2006; Mangano 2007; Royston 2015). This led to its withdrawal from many national markets in 2007 (FDA 2007). This data has, however, been revisited and reanalyzed, questioning the validity of the conclusions of the four studies (Howell 2013). Despite this, aprotinin remains unavailable or on a restricted license, for example for myocardial revascularization only, in some countries (Henry 2011).

Other haemostatic drugs

Desmopressin (DDVAP)

Desmopressin is a synthetic analogue of the human anti‐diuretic hormone, vasopressin. It increases the plasma levels of von Willebrand factor (vWF) two‐ to three‐fold by stimulating vWF release from endothelial cells. vWF plays an important role in platelet adhesion to wound sites, and thus early clot formation, so deficiency of vWF leads to bleeding tendencies. vWF also increases the availability of factor VIII, because factor VIII degrades rapidly if not complexed to vWF. Activated factor VIII is required in the enzyme cascade, which produces thrombin and fibrin. vWF deficiency is the most common clotting disorder and is present in about 1% of the population. Desmopressin is mainly used to treat coagulopathy caused either by deficiency of vWF or factor VIII (haemophilia A), but may also be used before procedures to treat reduced platelet adhesiveness due to drugs like aspirin, or from raised serum urea in the setting of severe renal impairment (Kim 2015).

Desmopressin is typically administered at a dose of 0.3 μg per kg subcutaneously or intravenously and takes approximately 30 minutes to reach peak effectiveness, and this effect lasts up to six to eight hours (Franchini 2007). Increases in vWF, factor VIII levels and in tissue plasminogen activator (tPA) if recurrent dosing is used can potentially increase the risk of arterial or venous thrombotic events; this is an important safety consideration (Franchini 2007; Kaufmann 2003). Desmopressin also results in release of nitric oxide from endothelial cells, which can cause vasodilation with symptoms of facial flushing, tachycardia, and hypotension (Kaufmann 2003). In rare cases, desmopressin administration may be associated with hyponatraemia and seizures, especially in young children (Smith 1989).

Prothrombin complex concentrate (PCC)

There are two main types of PCC. 3‐factor PCC contains blood clotting factors II, IX and X, whereas 4‐factor PCC also contains blood clotting factor VII, protein C, and protein S. PCC is a powder concentrate, extracted from human plasma and reconstituted prior to use, dosed at 25 to 50 units per kg. It is used for perioperative prophylaxis or treatment of severe bleeding in people treated with vitamin K antagonists, like warfarin, or in people with clotting factor deficiencies, whether inherited, for example haemophilia, or acquired, such as in severe liver disease (BNF 2019). Side effects include fever, high blood pressure and thromboembolism (migrating blood clots).

Recombinant factor VIIa (rFVIIa)

rFVIIa, also called NovoSeven, is a serine protease which catalyses conversion of factors IX and thrombin (X) into active forms. This increases the conversion of fibrinogen to fibrin by thrombin and promotes clot formation and propagation. It is currently licensed only for bleeding in people with a diagnosis of haemophilia, or severe uncontrolled haemorrhage, but is also used for prevention of haemorrhage in haemophiliacs undergoing invasive procedures like surgery (Simpson 2012). Studies have suggested an association with rFVIIa and arterial thromboembolic events (Levi 2010; Simpson 2012).

Factor XIII (FXIII)

FXIII, is a transglutaminase enzyme which cross‐links fibrin monomers between adjacent fibrin polymer strands to stabilise and strengthen the clot. It also acts to contract the clot into a more dense and insoluble unit (Ariëns 2002). FXIII treatment is currently indicated for congenital or acquired factor XIII deficiencies, identified with quantitative methods, and has been studied as an agent that can reduce bleeding in cardiac surgery (Muszbek 2008).

Fibrinogen concentrate

Fibrinogen is a plasma glycoprotein synthesised by the liver. Fibrinogen is the precursor to fibrin, but also helps platelets activate and aggregate by binding to the platelet’s GPIIb/IIIa receptor. Fibrinogen substitution is believed to normalize and improve the environment for clot formation by providing sufficient amounts of substrate and by enhancing the strength and speed of clot generation in people with depleted or dysfunctional fibrinogen (Nielsen 2005a; Nielsen 2005b). Within the context of cardiac surgery, systemic fibrinogen replacement is currently indicated for prophylaxis or treatment of bleeding in congenital and acquired deficiencies of fibrinogen that have been identified with quantitative methods (Bracey 2017). It has, however, been associated with small reduction in transfusions in a Cochrane Review of people with bleeding in elective and cardiac surgery, though without survival benefit (Wikkelsø 2013).

Internal topical agents (excludes surface dressings)

Internal topical application of drugs or biomaterials can be used as an adjunct to surgical control of bleeding, particularly where there are many microscopic bleeding vessels or raw tissue which cannot be surgically closed (Gabay 2013). A biomaterial is any substance that has been engineered to physically interact with biological tissue for a specific purpose (Park 2007). Topical agents include active drugs or clotting factors applied directly as a liquid, paste, foam or gel, or impregnated into biomaterials, or application of passive biomaterials which promote clotting through physical means (Vyas 2013). There are many agents available, and these have been classified as active, passive and combined haemostatic agents (Bracey 2017). They can also be classified as flowable, or non‐flowable, or fibrin and non‐fibrin sealant.

Active agents enhance enzyme pathways in clotting and include antifibrinolytic drugs, fibrin sealants or topical thrombin. Passive materials include collagens, porcine gelatins, regenerated oxidised cellulose and polysaccharide spheres. Passive synthetic sealants include cyanoacrylate, polyethylene glycol, and bovine serum albumin with glutaraldehyde. Combination agents include liquid gelatins with thrombin, and fibrin sealants with equine collagens. These diverse groups have the advantage of acting locally at the site of bleeding, potentially avoiding systemic side effects (Seyednejad 2008). The passive biomaterial and sealants may have the advantage of promoting clotting even in hypothermia or with deficits in normal clotting factors, as they operate independently from enzymatic biological clotting processes.

How the intervention might work

Antifibrinolytic drugs

Hyperfibrinolysis can contribute to catastrophic bleeding by preventing new clots forming as well as degrading formed clots. This is because fibrin degradation products interfere with platelet activation, adhesion and normal fibrin polymerisation, inhibiting normal coagulation. Additionally, the high level of free plasmin associated with hyperfibrinolysis also causes degradation of the fibrin precursor fibrinogen, reducing the substrate available for fibrin polymerisation. Prophylactic antifibrinolytic use is routinely used in all vascular surgery where excessive bleeding is anticipated (Chee 2016).

Other haemostatic drugs

Other haemostatic drugs are currently only recommended where a pre‐existing clotting factor deficiency has been identified with quantitative testing. There is a lack of well‐conducted studies to assess the impact of haemostatic drugs in people who may acquire perioperative deficits in clotting factors or have platelet function deficits due to perioperative medications. DDVAP may be of particular benefit in people with bleeding stemming from GPIIb/IIIa inhibitors and other antiplatelet medications (Raja 2006). rFVIIa is used off‐label for a variety of major surgeries, occasionally as prophylaxis, or more frequently in catastrophic haemorrhage after other options have failed to arrest bleeding. Its usefulness in reducing bleeding in surgery remains unproven (Simpson 2012). Analysis of rFVIIa usage in intractable bleeding in cardiothoracic surgery demonstrated a reduction in transfusion requirement, at the expense of a higher thrombotic event rate; it has not been determined, however, whether this translates into more favourable clinical outcomes (Omar 2015). Fibrinogen may be used during massive transfusion, but is not routinely used (Chee 2016). Therefore FXIII, rFVIIa and fibrinogen concentration may be used as a rescue treatment in severe bleeding rather than as prophylaxis due to their cost and risk profile.

Internal topical agents

Several trials show improved local haemostasis and reductions in overall blood use with topical agents, and there are theoretical advantages of localised treatments in terms of avoiding unwanted side effects (Vyas 2013). In people with abnormal clotting, however, local active treatments which rely on coagulation pathways to work may also have limited effect due to systemic coagulation derangement.

Why it is important to do this review

Bleeding and reoperation for bleeding are serious adverse outcomes, which are associated with increased mortality, complications, and risk of transfusion (Shaw 2013). Bleeding and the need for a red blood cell transfusion have also been shown to increase the duration of hospital stay and the costs associated with surgery, after taking into consideration confounding factors (Stokes 2011; Zbrozek 2015). The negative impact on outcomes associated with allogenic transfusion is observed even when a person only receives a transfusion of one or two units of red blood cells (Paone 2014; Paone 2018). These findings have recently been replicated in studies of major vascular surgery: after adjustment of major co‐variates, perioperative transfusion was associated with increased 30‐day mortality and morbidity (specifically myocardial infarction and pneumonia) in people undergoing major vascular surgery (Obi 2015). In lower limb bypass surgery, transfusion was associated with increased perioperative wound infection and graft thrombosis in a dose‐dependent fashion (Tan 2015). This has also been demonstrated in amputation surgery (Tan 2013). The particular blood product components transfused (red cells, platelet, fresh frozen plasma) may also impact outcome in AAA rupture surgery (Henriksson 2013). This study showed that the ratio of platelets and fresh frozen plasma to red cells increased from 0.8 to 0.9 during the study (1992 to 1999 versus 2000 to 2008), which was associated with improved survival.

Why even a small transfusion of red cells may be associated with poorer outcomes is not fully understood. It may be due to a mixture of pro‐inflammatory and anti‐inflammatory molecules within the transfusion, called transfusion‐related immunomodulation (TRIM) (Karsten 2018; Muszynski 2017; Youssef 2017). Other transfusion‐related adverse effects include incompatibility reactions, transfusion‐related acute lung injury (TRALI), and transfusion‐associated circulatory overload (TACO) (Harvey 2015; Maxwell 2006). In addition, transmission of infectious diseases (e.g. HIV, Hepatitis C, prion disease) remains a concern (Kiely 2017; Rerambiah 2014). This is particularly a concern in countries with higher prevalence of infectious diseases, or less robust screening capabilities, or both (Seo 2015; WHO 2017). Blood components, particularly platelets, can also have bacterial contamination that may cause sepsis in the recipient (Benjamin 2016; Makuni 2015; Morel 2013).

Adjuncts to reduce bleeding include prophylactic haemostatic drugs which alter coagulation and fibrinolysis. Tranexamic acid is probably the most frequently used at present, though aprotinin is re‐emerging after its withdrawal in the late 2000s (Royston 2015). These drugs may be given as a single dose, multiple doses or infusion, and before, during or after surgery, or by various different routes (e.g. topically onto a bleeding internal tissue, subcutaneously or intravenously). Audits of elective surgery show that there is poor uptake of pharmacological adjuncts to reduce bleeding (National Comparative Audit of Blood Transfusion). Barriers to optimal use may include not knowing which drug, drug combination, dose or timing is most effective. These factors are also important for establishing minimum effective doses and appropriate duration of exposure, so that other drug side effects are minimised. In order to select the most appropriate drug (or drug combination, dose, timing and route), the many different ways of giving these drugs should be compared; and this requires clarification and review of available evidence. Finally, this review will investigate the effect of antiplatelet/anticoagulant drug use and compare drug efficacy and safety in open and endovascular procedures, to establish any different performance of drugs in different circumstances (Berger 2012).

Objectives

To assess the efficacy and safety of anti‐fibrinolytic and haemostatic drugs and agents in reducing bleeding and transfusion in people undergoing major vascular surgery or vascular procedures with a risk of moderate or severe (> 500 mL) blood loss.

Methods

Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials (RCTs) in all languages. We will include cluster RCTs, if the analyses have accounted for clustering, or if we are able to adequately adjust for clustering (Hill 2018). If data are not available, we will contact the study authors; or if we are unable to obtain the necessary data, we will report the data in a narrative form with tables. We will include RCTs published as abstracts if they contain sufficient data, or if sufficient data is provided by the study authors on request. We will exclude studies with purely experimental laboratory outcomes (for example blood tests for inflammatory markers).

Types of participants

We will include adults (18 years or over) undergoing the following emergency, urgent and elective procedures.

• Open surgical repair (OSR) of aneurysm of the

  • abdominal aorta (AAA);

  • thoracic aorta (TAA);

  • thoraco‐abdominal aorta (TAAA);

  • iliac artery;

  • femoral artery; or

  • popliteal artery.

• Endovascular aneurysm repair (EVAR) of the

  • abdominal aorta (AAA);

  • thoracic aorta (TAA);

  • thoraco‐abdominal aorta (TAAA);

  • iliac artery;

  • femoral artery; or

  • popliteal artery.

• OSR or endovascular repair of dissection of the

  • abdominal aorta;

  • thoracic aorta; or

  • thoraco‐abdominal aorta.

• Open bypass surgery for peripheral arterial disease of the

  • aortic artery;

  • iliac artery;

  • femoral artery; or

  • popliteal artery

• Endovascular stenting for peripheral arterial disease of the

  • aortic artery;

  • iliac artery;

  • femoral artery; or

  • popliteal artery

• Major lower limb amputation for vascular disease

  • below knee;

  • above knee; or

  • hindquarter

We will include participants undergoing surgery with or without aortic cross‐clamping and with or without use of hypothermia. We will include participants undergoing open, modifications of open, and minimally invasive (e.g. laparoscopic) surgical approaches.

We will exclude procedures typically performed by or in conjunction with cardiac surgeons, such as those on the ascending aorta and aortic root, or those using coronary artery bypass grafting. These are the topic of a separate ongoing review entitled 'Drugs to reduce bleeding and transfusion in adults undergoing cardiac surgery; a systematic review and network meta‐analysis' (Beverly 2019).

We will exclude procedures associated with minimal bleeding and transfusion such as carotid procedures, arteriovenous fistulae formation for dialysis, varicose vein surgery and upper limb or digit amputations. We will exclude procedures typically performed by neurosurgeons such as repair of aneurysms or dissection of arteries in the head or neck.

We will exclude people with known inherited coagulation disorders, such as von Willebrand factor deficiency, haemophilia or hypofibrinogenaemia. This is because the clotting mechanisms that the drugs promote or interact with may be genetically absent, making response atypical.

For trials consisting of mixed populations of participants (e.g. including children, or including procedures other than those specified), only data from participants 18 years or over undergoing the specified procedures and without clotting disorders will be used. If the subgroup data required is not provided, we will exclude the trial if less than 80% of participants are eligible to be included in this review.

Types of interventions

We will include RCTs that compare one or more of the following interventions, compared to usual care, placebo, or each other.

• Tranexamic acid

• ε‐aminocaproic acid

• Aprotinin

• Desmopressin

• Prothrombin complex concentrate (PCC)

• Recombinant factor VII (rFVII)

• Factor XIII (FXIII)

• Fibrinogen concentrate

• Other topical agents, categorised as:

  • fibrin‐based agents;

  • thrombin‐based agents;

  • synthetic sealants;

  • passive biomaterials;

  • combination agents.

We will include studies using combinations of the above drugs. We will not exclude trials on the basis of the route, dose, timing, or frequency of drug administration. The comparison groups will be as defined by the study, which could be a control group using placebo, standard care, or one of the included drugs, if a second drug is being investigated.

Types of outcome measures

Primary outcomes

• Red cell transfusions (units per participant*) at up to 30 days post surgery

• All‐cause mortality at up to 30 days; and between 31 to 90 days

*If the red cell transfusion outcome is reported in millilitres, we will convert that into units, according to any local mean unit volume data given in the study, or as per the 'Guidelines for the Blood Transfusion Services in the UK' mean stated volume per unit of red cells of 280 ± 60 mL (JPAC 2013).

Secondary outcomes

• Risk of receiving any allogenic blood product at up to 30 days post surgery

  • Composite: packed red cells (PRC), fresh frozen plasma (FFP), platelets (PLTs)

  • Components: PRC, FFP, PLTs

• Risk of reoperation or repeat procedure for bleeding within 7 days

• Risk of a thrombotic/thromboembolic event

  • Composite: myocardial Infarction (MI), cerebrovascular attack (CVA), deep vein thrombosis (DVT), pulmonary embolus (PE) at up to 30 days and between 31 to 90 days

  • Components: MI at up to 30 days, CVA at up to 30 days, DVT at up to 90 days, PE at up to 90 days

• Risk of a serious adverse event (SAE**) at up to 30 days post surgery

• Length of hospital stay (days)

**We will use the International Conference on Harmonisation Good Clinical Practice definition of SAEs (ICH GCP 2018). Where studies report an SAE outcome using an alternative definition, we will record what definition of SAE is used and discuss suitability for inclusion in analysis with an expert panel, prior to extracting outcome data.

We will comment on any cost data, if presented, in a narrative form (Ryan 2016). Cost information will provide useful additional information, but will not be a formal economic evaluation.

Where a published report does not appear to report on one or more of these primary or secondary outcomes, we will access the trial protocol and contact the trial authors for data in case an outcome measure was collected in the protocol but was not reported. We will include as part of the narrative in the systematic review any relevant trials which measured outcomes but did not report the data in a usable format.

Search methods for identification of studies

Electronic searches

Searches will use a combination of MeSH and free text terms and we will carry them out from database inception to the present, irrespective of language or publication status.

The Information Specialist will search the following databases for relevant trials.

• Cochrane Central Register of Controlled Trials (CENTRAL) via the Cochrane Register of Studies Online (CRSO)

• MEDLINE (Ovid MEDLINE® Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE® Daily and Ovid MEDLINE®) (1946 onwards)

• Embase Ovid (from 1974 onwards)

• CINAHL EBSCO (from 1982 onwards)

• Transfusion Evidence Library (1950 to present)

The Information Specialist will search the following trials registries.

• World Health Organization International Clinical Trials Registry Platform (who.int/trialsearch)

• ClinicalTrials.gov

We will apply the Cochrane sensitivity‐ and precision‐maximizing RCT filter to Ovid MEDLINE (Lefebvre 2011), and adaptations of it to Ovid Embase and Ovid Emcare, in combination with a systematic review filter (to include systematic reviews to allow manual screening for additional citations; see Searching other resources) based on the Scottish Intercollegiate Guidelines Network (SIGN) filter (www.sign.ac.uk/methodology/filters.html). The Information Specialist has devised a MEDLINE draft search strategy for RCTs which is displayed in Appendix 1. We will use this as the basis for search strategies for the other databases listed.

Searching other resources

We will check reference lists of all included studies and any relevant systematic reviews we identify for additional references to trials. We will also examine any relevant retraction statements and errata for included studies. We will contact authors in the field to help identify additional studies (Lefebvre 2011).

Data collection and analysis

We will conduct and report the review in accordance with guidelines in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019), using any latest chapter updates as available; and also in accordance with the PRISMA checklist extension for network meta‐analyses (Hutton 2015).

Selection of studies

Two review authors (AB, GO) will use Covidence software to screen electronically derived citations and abstracts of papers identified by the review search strategy for relevance. We will exclude any studies that are clearly irrelevant at this stage. We will retrieve the full text of references when we cannot assess eligibility by title and abstract alone. We will translate studies reported in non‐English‐language journals before assessment. We will exclude studies not meeting eligibility criteria from further analysis and detail these in the 'Characteristics of excluded studies' table.

At all stages the two review authors will try to resolve any disagreements by discussion but, if necessary, they will consult with a third review author (LE). We will record the reasons why studies failed to meet the inclusion criteria and display the results of the search in a PRISMA flow chart (Hutton 2015).

If we include any multi‐arm studies with more than two relevant arms, we will only add it once to the 'Characteristics of included studies' table, to avoid ‘double counting’ the study in terms of the number of RCTs, and to prevent the study appearing twice in the 'Risk of bias' figures.

Data extraction and management

Two review authors (AB, GO) will independently undertake data extraction from included studies. We will design, pilot and modify data extraction forms. We anticipate there are many potential options for dose, route and timing within each drug intervention.

We will consider the following data for extraction from each study, dependent on feasibility and performance in pilot.

General information

Country of study, single or multi‐centre, funding source, publication type (abstract/full text/protocol), trial registration and timing (prospective or retrospective), year of publication.

Trial details

Trial design, funding, location, setting, number of centres, number of treatment arms, intention‐to‐treat analysis, power calculation and whether reached, treatment allocation method, randomisation, blinding, total number recruited, total number randomised, total number analysed in each study group, dropout rate, participant inclusion and exclusion criteria, antiplatelet and anticoagulant cessation protocol, transfusion strategy, comparability of groups according to 'participants characteristics', length of follow‐up, stopping rules, thrombotic event definition, SAE definition.

Characteristics of participants

Age, sex, ethnicity, weight, pre‐operative antiplatelet and anticoagulant medication (including washout period).

Characteristics of surgery

Type of vascular operation, risk stratification, urgency of surgery (e.g. elective, non‐elective, mixed, not stated), surgical duration, aortic cross‐clamp use, aortic cross‐clamp duration, use of hypothermia, mean minimum temperature, percentage in each arm dropping out (with reasons), percentage in each arm lost to follow‐up.

Characteristics of intervention

Number of arms, description of intervention and comparison arms, description of control arms (including placebo, usual care etc.), intervention(s) given, route of administration of intervention, timing of intervention, methods of dosing (e.g. standard, dose/kg, dose categories), dose, dose delivery (single bolus, multiple bolus, infusion).

Grouping interventions into treatment nodes for data synthesis

The included studies may use a range of different doses, routes or timings of interventions. Therefore we will undertake an intervention taxonomy process to identify if any interventions are similar enough to cluster together into nodes for a subsequent network meta‐analysis. We will make a list of all the different types of each intervention (dose, route, timing, bolus, repeat bolus, infusion) encountered across the included studies. We will then consult an independent, expert panel to determine potential groupings of the interventions which can be usefully clinically compared. We will include anaesthetic, surgical, haematological, pharmacological and statistical expertise within this panel review.

We will do this after extracting participant, methodology and intervention data but prior to extracting outcome data, to avoid introducing bias. This will therefore be blinded in so far as authors and the panel will not have access to the results of data extraction at this point, though some may have some incidental familiarity with the literature. This two‐stage approach was devised with experts in the area of complex systematic reviews at the National Institute of Healthcare Research (UK) Complex Review Support Unit. In this way we aim to ensure that any network meta‐analysis of interventions will be meaningful, relevant and feasible.

Two review authors (AB, GO) will then independently extract outcome data.

Assessment of risk of bias in included studies

For the systematic review and network meta‐analysis, two review authors (GO, AB) will independently assess the risk of bias using the Cochrane 'Risk of bias' tool (RoB) (Higgins 2017). We will resolve any disagreements by discussion or by consulting a third review author (LE).

For the network meta‐analysis we will use the corresponding RoB tool for network meta‐analysis, called Confidence in Network Meta Analysis (CiNeMA 2017).

We will conduct the review according to this published protocol and report any deviations from it in the 'Differences between protocol and review' section of the systematic review.

Measures of treatment effect

We will express measures of treatment effect using the criteria laid out by Cochrane for dichotomous outcomes and continuous outcomes.

For dichotomous outcomes, we will record the number of events and total number of participants in treatment and control groups. For continuous outcomes, we will record the mean, standard deviation and total number of participants in both the treatment and control groups.

For dichotomous variables, we will express the results as odds ratios (ORs) with 95% confidence intervals (CIs). Where the number of observed events is small (less than 5% of sample per group), and where trials have balanced treatment groups, we plan to report the Peto's OR with 95% CIs (Deeks 2017).

For continuous variables using the same scale, we will express the results as mean difference (MD) with 95% CIs (Takeshima 2014). For continuous outcomes measured with different scales, we will present the standardised mean difference (SMD) with 95% CIs. We will interpret the SMD in line with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019). We will present SMDs as standard deviation units and interpret as follows: 0.2 represents a small effect; 0.5 a moderate effect; and 0.8 a large effect (Cohen 1988). We will report these measures of treatment effects alongside associated 95% CIs. We are aware that certain outcomes, for example red cell transfusions, could be reported with different units (mL, mL/kg, units). Therefore, if we are unable to convert all these measures to the desired unit (e.g. units of packed red cells) we will conduct separate meta‐analysis.

If available, we plan to extract and report hazard ratios (HRs) for time‐to‐event data (mortality or time in hospital). If HRs are not available, we will make every effort to estimate as accurately as possible the HR using the available data and a purpose‐built method based on the Parmar and Tierney approach (Parmar 1998; Tierney 2007). If sufficient studies provide HRs, we plan to use these in favour of other reported treatment effects in the meta‐analysis, otherwise we will perform a separate meta‐analysis for all types of reported treatment effects, for example ORs. If the events are rare and the follow‐up times are similar we will consider the perceived similarity of ORs, risk ratios (RRs) and HRs.

We will produce narrative descriptions of skewed data, reported as medians and interquartile ranges. If we cannot synthesize the data, we will provide a descriptive narrative summary and tables with the available information. When we cannot report available data in any of the formats described above we will provide a narrative report and, when appropriate, present the data in tables.

Unit of analysis issues

We will consider participants as the unit of analysis (McKenzie 2016).

Cluster randomised trials

We will consider and report on the impact of including any cluster trials in our analysis. We will respect the appropriate unit of analysis (e.g. hospitals), and ensure appropriate treatment of data from cluster RCTs. We will extract and report direct estimates of the effect measure (e.g. OR with 95% CIs) from an analysis if it accounts for the clustered design. If appropriate analyses are not available, we will make every effort to approximate the analysis following the recommendations in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019).

Studies with multiple treatment groups

In pairwise meta‐analyses, we will treat trials with multiple treatment group comparisons as individual, independent two‐arm studies. To avoid duplicate data in pairwise meta‐analysis, multiarm studies will either be combined if it is clinically appropriate, or we will make appropriate adjustments; for example, dividing the sample size of a group for a continuous outcome.

We will include all different interventions in the network meta‐analysis, if participant populations are comparable, regardless of how many arms there are in the study. The control group will act as a node in the network meta‐analyses, which will help with indirect analyses and formation of a hierarchy of interventions.

Dealing with missing data

We will handle missing data according to the methods described in Chapter 16 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019).

We will contact investigators or study sponsors in order to verify key study characteristics and obtain missing numerical outcome data where possible (e.g. when a study is identified as abstract only). Where possible, we will use the Review Manager 5 (RevMan 5) calculator to calculate missing standard deviations using other data from the trial, such as confidence intervals (Review Manager 2014). Where this is not possible, and we think the missing data introduce serious bias, we will explore the impact of including such studies in the overall assessment of results by a sensitivity analysis (Higgins 2019).

Assessment of heterogeneity

If we consider the clinical and methodological characteristics of individual studies to be sufficiently homogenous, we will combine the data to perform a meta‐analysis (Deeks 2017). If excessive heterogeneity is present, we will explore the data and try to understand the underlying reasons. In standard pairwise meta‐analyses, we will estimate the heterogeneity variances for each pairwise comparison. In network meta‐analysis, we will assume a common estimate for the heterogeneity variance across the different comparisons.

Measures and tests for heterogeneity

During initial data extraction, we will assess if clinical and methodological heterogeneity are present by looking at trial and person characteristics across all included trials.

We will use STATA to assess statistical heterogeneity within each pairwise comparison using the I² statistic and its 95% CI, which measures the percentage of variability that cannot be attributed to random error (I² > 50% = moderate heterogeneity; I² > 80% = considerable heterogeneity) (Stata).

The assessment of statistical heterogeneity in the entire network will be based on the magnitude of the heterogeneity variance parameter (Tau²) estimated from the network meta‐analysis models. We will assume a random‐effects model, but we will test this assumption with a likelihood ratio test comparing random‐effects and fixed‐effect models (Borenstein 2010). We will compare the magnitude of the heterogeneity variance with previously suggested empirical distributions (Rhodes 2015). We will also estimate a total I² value for heterogeneity in the network and estimate prediction intervals for all relative effects. We will explore potential causes of heterogeneity by subgroup analyses if possible (Dias 2018).

Assessment of transitivity across treatment groups

We will assess the assumption of transitivity by comparing the distribution of the potential treatment effect modifiers across the different pairwise comparisons. We will also epidemiologically evaluate the assumption of transitivity by comparing the clinical and methodological characteristics of sets of studies grouped within the network (Chaimani 2018; Jansen 2013).

Assessment of reporting biases

If we are able to pool more than 10 trials for pairwise meta‐analyses, we will create and examine a funnel plot to explore possible small‐study biases for the primary outcomes and use a formal statistical test for asymmetry (Egger 1997). For investigating small‐study effects, an alternative Harbord or Sterne test will be used (Harbord 2006; Sterne 2002).

Data synthesis

Methods for direct treatment comparisons

We will perform the statistical pairwise meta‐analysis using RevMan 5 software (Review Manager 2014).

We will use a random‐effects model unless there are rare events in which case we will use the Peto's OR which is only available in a fixed‐effect model. For continuous outcomes we will use the inverse variance method and meta‐analyse mean difference; or if there are different scales used, we will use SMD. For time‐to‐event data (reported as HRs) we will use the generic inverse variance method. We will also use this if we have to combine appropriate cluster trials.

If we use random‐effects analyses, we will present the results as the average treatment effect with its 95% CI, and the estimates of Tau² or I².

We will present all data that cannot be included in meta‐analyses in tables.

Methods for network meta‐analysis

We will undertake meta‐analyses only where this is meaningful: that is, if the participants and the underlying clinical question are not overly heterogeneous (Salanti 2008). We will create network diagrams for each outcome to check if the network is connected with direct or indirect comparisons. The expert panel together with the network diagrams will guide us towards grouping the interventions, comparisons, routes, doses and timings into appropriate nodes as described above.

We will perform network meta‐analysis in Stata using the method of multivariate meta‐analysis that treats the different comparisons in studies as different outcomes (Stata; White 2012). We will perform this analysis using a network package with the 'mvmeta' command (White 2011); and we will present the results using the 'network graphs' package in Stata (Chaimani 2013; Chaimani 2015). To evaluate the assumption of transitivity, we will compare the distribution of the potential effect modifiers across the available direct comparisons. A list of pre‐specified effect modifiers is in the 'Treatment effect modifiers' section below. We will assess inconsistency between the evidence from direct comparison and indirect comparison for interventions included in the network analysis, which is an additional method to assess transitivity of treatment effect. We will consider using both loop and global approaches as appropriate (Dias 2013).

Relative treatment ranking

We plan to obtain a hierarchy of the competing interventions in the network and to present these with probability of different rankings tables (Mbuagbaw 2017).

Subgroup analysis and investigation of heterogeneity

Treatment effect modifiers

We plan to investigate the population characteristics potentially acting as treatment effect modifiers by carrying out the following subgroup analyses.

• Endovascular versus open surgery

• Perioperative antiplatelet and anticoagulant therapy

• Aortic cross‐clamp use

• Hypothermia use

We will use the formal test for subgroup differences in RevMan 5 (Review Manager 2014), and base our interpretation on this. We will use this for evaluation of the transitivity assumption. We will conduct subgroup analysis on both the pairwise meta‐analysis and the network meta‐analysis.

Sensitivity analysis

We plan to carry out the following sensitivity analyses, to test whether key methodological factors or decisions have affected the main results.

• Only including studies with a low risk of bias

• Broader versus narrower groupings of interventions (on the basis of similar dosing, timings and routes)

Our sensitivity analysis, 'including only studies with a low risk of bias', will be based on the overall risk of bias assessment. We will conduct sensitivity analysis on both the pairwise meta‐analysis and the network meta‐analysis.

Summary of findings and assessment of the certainty of the evidence

Pairwise analysis

We plan to extract study data, format our comparisons in data tables and prepare a 'Summary of findings' table for the paired meta‐analysis before writing the results and conclusions of our review.

We will create a 'Summary of findings' table for each intervention using the following outcomes: proportion of participants receiving any allogenic blood product; all‐cause mortality; mean number of red cell transfusions; proportion of participants experiencing reoperation for bleeding; thrombotic/thromboembolic events; serious adverse events; and length of hospital stay.

We will use the five GRADE considerations (study limitations; consistency of effect; imprecision; indirectness; and publication bias) to assess the certainty of a body of evidence as it relates to the studies which contribute data to the meta‐analyses for the pre‐specified outcomes. We will use methods and recommendations described in Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2017) using GRADEpro software (GRADEpro GDT). We will justify all decisions to downgrade the certainty of evidence using footnotes; and we will make comments to aid the reader's understanding of the review where necessary.

Two review authors (AB, GO) will independently make judgements about evidence certainty, with disagreements resolved by discussion or by involving a third review author (LE). Judgements will be justified, documented and incorporated into reporting of results for each outcome.

We have included an example 'Summary of findings' table for the comparison 'Tranexamic acid compared with placebo for elective open repair' (Table 1).

1. Example Summary of Findings table.

e.g. Tranexamic acid compared with placebo for elective open repair
Patient or population: e.g. people aged 18 and over with abdominal aortic aneurysm Settings: hospital Intervention: e.g. tranexamic acid, 1 g to 2 g, IV, at induction, not repeated Comparison: e.g. placebo
Outcomes	Anticipated absolute effects * (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with placebo	Risk with tranexamic acid
Red cell transfusions (mean units per participant, follow‐up)
All‐cause mortality (follow‐up)
Risk of receiving any allogenic blood product (follow‐up)
Risk of undergoing reoperation or repeat procedure for bleeding (follow‐up)
Risk of experiencing a thrombotic/thromboembolic event (composite of MI, CVA, DVT, PE) (follow‐up)
Risk of experiencing SAEs (follow‐up)
Mean length of hospital stay (days, follow‐up)
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CVA: cerebrovascular attack; DVT: deep vein thrombosis; IV: intravenous; MI: myocardial infarction; PE: pulmonary embolus; RR: risk ratio; SAEs: serious adverse events
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: we are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: we have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Network meta‐analysis (NMA)

Prior to employing NMA, we will determine its suitability. We will then use GRADEpro GDT and CiNeMA to assess and present the certainty of evidence within the NMA (CiNeMA 2017;Nikolakopoulou 2020; Papakonstantinou 2020). We will use the same five GRADE domains as for pairwise analysis. We will present the summary of findings as intervention hierarchies for each given outcome, rather than a 'Summary of findings' table (Mbuagbaw 2017). The outcomes selected will be the same as those in the 'Summary of findings' table for pairwise analysis, and deviation from this will be stated in the 'Differences between protocol and review' section of the final review.

History

Protocol first published: Issue 7, 2020

Appendices

Appendix 1. MEDLINE search strategy

1. Vascular Surgical Procedures/

2. Endarterectomy/

3. exp Endovascular Procedures/

4. Axillofemoral Bypass Grafting/

5. Embolectomy/

6. Limb Salvage/

7. exp Amputation/

8. exp Thrombectomy/

9. Vascular Grafting/

10. Arteriovenous Shunt, Surgical/

11. Blood Vessel Prosthesis Implantation/

12. Vascular Diseases/su

13. Femoral Artery/su or Iliac Artery/su or Popliteal Artery/su

14. exp Aneurysm/su

15. exp Aortic Diseases/su

16. exp Arterial Occlusive Diseases/su

17. exp Arteriovenous Malformations/su

18. Diabetic Angiopathies/su

19. exp Peripheral Vascular Diseases/su

20. exp Spinal Cord Vascular Diseases/su

21. Vascular System Injuries/su

22. exp Venous Insufficiency/su

23. or/1‐22

24. Radiology, Interventional/

25. (interventional radiolog* or surgical radiolog*).tw,kf.

26. ((vascular or vessel* or aneurysm or aortic or aorta or AAA or TAA or TAAA or aortofemoral or artery or arterial or interarterial or arterioarterial or atheroma or carotid or vein or venous or endovascular* or intravascular*) adj6 (surg* or operat* or reconstruct* or repair* or resection* or repair* or bypass* or graft* or homograft* or transplant* or bypass* or transpos* or implant* or reimplant* or clamp* or ligat* or crossover* or cross‐over*)).tw,kf.

27. ((endovascular or intravascular or laparoscop* or angioscop*) adj3 (surg* or operat* or reconstruct* or repair* or resection* or repair* or bypass* or graft* or homograft* or transplant* or bypass* or transpos* or implant* or reimplant* or procedure*)).tw,kf.

28. (bypass graft* or by‐pass graft* or bypass surgery or by‐pass surgery or enarterectom* or (minimally invasive adj1 surg*) or teflon graft* or dacron graft*).tw,kf.

29. (arterial dilation* or endoluminal repair*).tw,kf.

30. ((venous or arterial) adj3 catheteri?ation).tw,kf.

31. ((aort* or iliac or femoral or popliteal or femoropop* or fempop* or crural) adj3 (surg* or operat* or bypass or graft or reconstruct* or revascular*)).tw,kf.

32. (angiosurg* or aneurysmectom* or aneurysm clipping* or aortopexy or aortoplast* or arteriotom* or arterioplast* or artery plast* or artery stripping or venostom* or portacaval anastomoses or revasculari?ation or devasculari?ation).tw,kf.

33. ((femoral* or iliac* or aorta* or aortic* or infrapopliteal or popliteal or infra‐pop*) adj3 (stent* or angioplast*)).tw,kf.

34. (endovascular adj6 (dissection* or stent*)).tw,kf.

35. (EVAR or FEVAR or TEVAR or embolectom* or thrombectom* or endarterectom* or thromboendarterectom* or thromboembolectom* or atherectom*).tw,kf.

36. (mechanical* adj3 (thrombolysis or clot disruption*)).tw,kf.

37. (axillo bifemoral bypass graft* or axillo femoral bypass graft* or axillobifemoral bypass graft* or axillofemoral bypass graft*).tw,kf.

38. ((ilio femor* or ilio femor* or infrarenal or infra‐renal or infrainguinal or infra‐inguinal) adj (bypass* or by‐pass*)).tw,kf.

39. ((pedal or tibial) adj3 (angio* or bypass* or by‐pass*)).tw,kf.

40. ((lower limb* or knee* or leg* or foot or feet or lower extremit* or hindquarter*) adj5 (salvag* or saving or save* or angioplast* or graft* or bypass* or by‐pass* or revascula* or reconstruct* or amputat*)).tw,kf.

41. ((blood vessel* or vascular or endovascular or vein or venous or arterial or artery or arteriovenous or arterio‐venous or brescia cimino or venoarterial or veno‐arterial or arterioportal) adj3 (anastomosis or shunt*)).tw,kf.

42. or/24‐41

43. 23 or 42

44. Antifibrinolytic Agents/

45. Tranexamic Acid/

46. Aminocaproic Acid/

47. (antifibrinolytic* or anti‐fibrinolytic* or antifibrinolysin* or antiplasmin* or plasmin inhibitor* or tranexamic or tranhexamic or cyclohexanecarboxylic acid* or amcha or trans‐4‐aminomethyl‐cyclohexanecarboxylic acid* or t‐amcha or amca or "kabi 2161" or transamin or amchafibrin or anvitoff or spotof or cyklokapron or cyclo‐F or femstrual or ugurol or aminomethylcyclohexanecarbonic acid or aminomethylcyclohexanecarboxylic acid or AMCHA or amchafibrin or amikapron or aminomethyl cyclohexane carboxylic acid or aminomethyl cyclohexanecarboxylic acid or aminomethylcyclohexane carbonic acid or aminomethylcyclohexane carboxylic acid or aminomethylcyclohexanecarbonic acid or aminomethylcyclohexanecarboxylic acid or aminomethylcyclohexanocarboxylic acid or aminomethylcyclohexanoic acid or amstat or antivoff or caprilon or cl?65336 or cl65336 or cyclocapron or cyclokapron or cyklocapron or cyklokapron or exacyl or frenolyse or fibrinon or hemostan or hexacapron or hexakapron or kalnex or lysteda or rikaparin or ronex or theranex or tranexam or tranexanic or tranexic or trans achma or transexamic or trenaxin or TXA or (fibrinolysis adj2 inhibitor*)).tw,kf.

48. (Agretax or Bio‐Stat or Capiloc or Capitrax or Clip Inj or Clot‐XL or Clotawin‐T or Coastat or Cuti or Cymin or Dubatran or Espercil or Examic or Existat or Extam or Fibran or Gynae‐Pil or Hemstate or Kapron or Menogia or Monitex or Nestran or Nexamic or Nexi‐500 or Nexmeff or Nicolda or Nixa‐500 or Pause or Rheonex or Sylstep TX or Synostat or T‐nex or T Stat or T Stat or Tanmic or Temsyl‐T or Texakind or Texanis or Texapar or Texid or Thams or Tonopan or Traklot or Tramic or Tramix or Tranarest or Trance Inj or Tranecid or Tranee or Tranemic or Tranex or Tranexa or Tranfib or Tranlok or Transtat or Transys or Transcam or Tranxi or Trapic or Traxage or Traxamic or Traxyl or Trenaxa or Trexamic or Trim Inj or Tx‐1000 or Tx 500 or Wistran or X‐Tran or Xamic).tw,kf.

49. (6‐aminohexanoic or amino?caproic or amino?hexanoic or amino caproic or amino‐caproic or amino‐n‐hexanoic or cy‐116 or cy116 or lederle or acikaprin or afibrin or amicar or caprocid or capracid or capramol or caprogel or caprolest or caprolisin* or caprolysin* or capromol or epsikapron or hemocaprol or caproamin or EACA or caprolest or capralense or hexalense or hamostat or hemocid or cl 10304 or cl10304 or ecapron or ekaprol or epsamon or epsicaprom or epsicapron or epsilcapramin or epsilon amino caproate or epsilon aminocaproate or epsilonaminocaproic or epsilonaminocapronsav or etha?aminocaproic or ethaaminocaproich or emocaprol or hepin or ipsilon or jd?177or neocaprol or nsc?26154 or resplamin or tachostyptan).tw,kf.

50. or/44‐49

51. Aprotinin/

52. (antagosan or antilysin* or aprotimbin or apronitin* or aprotinin* or bayer a128 or contrical or contrycal or contrykal or dilmintal or frey inhibitor or kontrycal or Kunitz inhibitor or gordox or haemoprot or kallikrein‐trypsin inactivator).tw,kf.

53. (iniprol or kontrikal or kontrykal or kunitz pancreatic trypsin inhibitor or midran or pulmin or tracylol or trascolan or trasilol or tra?ylol or traskolan or zymofren or pancreas antitrypsin or protinin or riker 52g or Rivilina zymofren).tw,kf.

54. or/51‐53

55. Factor VIIa/

56. (factor viia or factor 7a or rfviia or fviia or novoseven* or novo seven* or aryoseven or acset or eptacog* or proconvertin).tw,kf.

57. ((activated adj2 factor seven) or (activated adj2 factor vii) or (activated adj3 rfvii) or (activated adj2 fvii)).tw,kf.

58. (factor seven or factor vii or factor 7).ti.

59. 55 or 56 or 57 or 58

60. Fibrinogen/ad, ae, de, sd, tu, th, to

61. *Fibrinogen/

62. (fibrinogen concentrate* or factor I or Haemocomplettan* or Riastap* or Fibryga* or Fibryna*).tw,kf.

63. 60 or 61 or 62

64. Deamino Arginine Vasopressin/

65. (desmopressin* or vasopressin deamino or D‐amino D‐arginine vasopressin or deamino‐8‐d‐arginine vasopressin or vasopressin 1‐desamino‐8‐arginine or desmotabs or DDAVP or desmogalen or adin or adiuretin or concentraid or d‐void or dav ritter or deamino 8 dextro arginine vasopressin or deamino 8d arginine vasopressin or deamino dextro arginine vasopressin or deaminovasopressin or defirin or defirin melt or desmirin or desmomelt or desmopresina or desmospray or desmotab* or desurin or emosint or enupresol or minirin or minirinette or minirinmelt or minrin or minurin or miram or nictur or noctisson or nocturin or nocutil or nordurine or novidin or nucotil or octim or octostim or presinex or stimate or wetirin).tw,kf.

66. 64 or 65

67. exp Factor XIII/

68. (factor XIII or fXIII or fibrin stabili?ing factor* or Tretten* or Catridecacog).tw,kf.

69. 67 or 68

70. exp Tissue Adhesives/

71. *Adhesives/

72. Collagen/tu

73. Thrombin/tu

74. Gelatin/tu

75. Gelatin Sponge, Absorbable/

76. ((fibrin* or collagen or cellulose or gelatin or gel or thrombin* or albumin or hemostatic* or haemostatic*) adj3 (glu* or seal* or adhesive* or topical* or local* or matrix or matrices or spong* or fleece* or foam* or scaffold* or patch* or sheet* or bandag* or aerosol* or dressing* or paste or powder*)).tw,kf.

77. ((nonfibrin* or non‐fibrin* or synthetic* or non‐biological* or nonbiological* or biological*) adj3 (glue* or seal* or adhesive*)).tw,kf.

78. (surgical* adj3 (glue* or sealant* or adhesive*)).tw,kf.

79. ((fibrin* or collagen or cellulose or gelatin or thrombin) adj3 (hemosta* or haemosta*)).tw,kf.

80. (8Y or Aafact or Actif‐VIII or Advate or Artiss or Bioglue or Biocol or Collaseal or Omrixil or Transglutine or Raplixa or Evarrest or Aleviate or Alphanate or Amofil or Beriate or Beriplast or Biostate or Bolheal or Cluvot or Conco‐Eight‐HT or Crosseel or Crosseal or Crosseight or Emoclot or Evarrest or Evicel or Factane or Fanhdi or Fibrogammin P or Green VIII or Green VIII Factor or Greengene or Greenmono or Greenplast or Haemate or Haemate P or Haemate P or Haemate P500 or Haemate‐P or Haemoctin or Haemoctin SDH or Haemoctin‐SDH or Hemaseel or Hemaseal or Hemofil M or Hemoraas or Humaclot or Humafactor‐8 or Humate‐P or Immunate or Innovate or Koate or Koate‐DVI or Kogenate Bayer or Kogenate FS or Monoclate‐P or NovoThirteen or Octafil or Octanate or Octanate or Optivate or Quixil or Talate or Tisseel or Tisseal or Tissel or Tissucol or Tricos or Vivostat or Voncento or Wilate or Wilnativ or Wilstart or Xyntha).tw,kf.

81. (Glubran or Gluetiss or Ifabond or Indermil or LiquiBand or TissuGlu).tw,kf.

82. (Evithrom or Floseal or Hemopatch or Gel‐Flow or Gelfoam or Gelfilm or Recothrom or Surgifoam or Surgiflo* or "rh Thrombin" or Thrombi‐Gel or Thrombi‐Pad or Thrombin‐JMI or Thrombinar or Thrombogen or Thrombostat).tw,kf.

83. (porcine gelatin or bovine collagen or bovine gelatin or nu‐knit or arista or hemostase or vita sure or thrombin‐jmi or thrombinjmi or avicel or vivagel or lyostypt or tabotamp or arterx or omnex or veriset).tw,kf.

84. (polysaccharide adj (sphere* or hemostatic powder)).tw,kf.

85. *Chitosan/

86. *Polyethylene Glycols/

87. *Hydrogel, Polyethylene Glycol Dimethacrylate/

88. Polyurethanes/ad, ae, pd, tu, to

89. ((polymer‐derived elastic* or polymer tissue adhesive* or elastic hydrogel* or glutaraldehyde or PEG‐based or polyurethane‐based tissue or polyethylene glycol* or polyvinyl alcohol‐based tissue or PVA‐based tissue or natural biopolymer* or polypeptide‐based or protein‐based or polysaccharide‐based or chitosan or poliglusam or cyanoacrylic or cyanoacrylate or cyacrin or dextran‐based or chondroitin sulfate‐based or mussel‐inspired elastic* or glycol hydrogel or polymer‐based) adj3 (glu* or seal* or adhesive* or topical* or local* or matrix or matrices or spong* or fleece* or foam* or scaffold* or patch* or sheet* or bandag* or aerosol* or dressing* or paste* or powder*)).tw,kf.

90. Cellulose, Oxidized/

91. (absorbable cellulose or resorbable cellulose or oxidi?ed cellulose or carboxycellulose or oxycellulose or cellulosic acid or oxycel or oxidi?ed regenerated cellulose).tw,kf.

92. (BioGlue or Progel or Duraseal or Coseal or FocalSeal or ADAL‐1 or AdvaSeal or Pleuraseal or Angio‐Seal or Avitene or Instat or Helitene or Helistat or TDM‐621 or Dermabond or Tissueseal or PolyStat or Raplixa or Spongostan or Surgicel or Surgilux or Tachosil or Traumstem).tw,kf.

93. (collagen‐thrombin or thrombin‐collagen or gelatin‐fibrinogen or fibrinogen‐gelatin or gelatin‐thrombin or thrombin‐gelatin or fibrinogen‐thrombin or thrombin‐fibrinogen or collagen‐fibrinogen or fibrinogen‐collagen or microfibrillar collagen or CoStasis or "GRF Glue" or GR‐Dial or Algosterile or TraumaStat or HemCon or ChitoFlex or Celox or QuikClot or WoundStat or Vitagel or TachSeal or TachoComb or Cryoseal).tw,kf.

94. or/70‐93

95. exp Waxes/

96. (bonewax* or bone wax* or bone putty or hemasorb or ostene).tw,kf.

97. 95 or 96

98. Blood Coagulation Factors/

99. (prothrombin adj5 (complex* or concentrate*)).tw,kf.

100. (PCC* or 3F‐PCC* or 4F‐PCC* or Beriplex* or Feiba* or Autoplex* or Ocplex* or Octaplex* or Kcentra* or Cofact or Prothrombinex* or "Proplex‐T" or Prothroraas* or Haemosolvex* or Prothromplex* or "HT Defix" or Facnyne* or Kaskadil* or Kedcom* or Confidex* or PPSB or Profil?ine* or Pronativ* or Proplex* or Prothar* or ProthoRAAS* or Protromplex* or "Pushu Laishi" or "Uman Complex").tw,kf.

101. or/98‐100

102. (((haemosta* or hemosta* or antihaemorrhag* or antihemorrhag* or anti haemorrhag* or anti‐hemorrhag*) adj5 (drug* or agent* or treat* or therap*)) or ((coagulat* or clotting) adj factor*)).tw,kf.

103. 50 or 54 or 59 or 63 or 66 or 69 or 94 or 97 or 101 or 102

104. 43 and 103

105. Meta‐Analysis.pt.

106. Systematic Review.pt.

107. ((meta analy* or metaanaly*) and (trials or studies)).ab.

108. (meta analy* or metaanaly* or evidence‐based).ti.

109. ((systematic* or evidence‐based) adj2 (review* or overview*)).tw,kf.

110. (evidence synthes* or cochrane or medline or pubmed or embase or cinahl or cinhal or lilacs or "web of science" or science citation index or scopus or search terms or literature search or electronic search* or comprehensive search* or systematic search* or published articles or search strateg* or reference list* or bibliograph* or handsearch* or hand search* or manual* search*).ab.

111. Cochrane Database of systematic reviews.jn.

112. (additional adj (papers or articles or sources)).ab.

113. ((electronic* or online) adj (sources or resources or databases)).ab.

114. (relevant adj (journals or articles)).ab.

115. or/105‐114

116. Review.pt.

117. exp Randomized Controlled Trials as Topic/

118. selection criteria.ab. or critical appraisal.tw.

119. (data adj (abstract* or extract* or analys*)).ab.

120. exp Randomized Controlled Trial/

121. or/117‐120

122. 116 and 121

123. 115 or 122

124. Randomized Controlled Trial.pt.

125. Controlled Clinical Trial.pt.

126. (placebo or randomly or groups).ab.

127. (randomi* or trial).tw,kf.

128. exp Clinical Trial as Topic/

129. or/123‐128

130. exp animals/ not humans/

131. 129 not 130

132. 104 and 131

Contributions of authors

AB: protocol development
GO: protocol development
CD: search strategy and protocol development
LE: content expert, and protocol development
NJW: expert advice on network meta‐analysis

Sources of support

Internal sources

• National Institute of Healthcare Research (NIHR), UK

  NIHR grant to NHS Blood and Transplant

• NIHR Biomedical Research Centre at University Hospitals Bristol NHS Foundation Trust and the University of Bristol, UK

External sources

• Chief Scientist Office, Scottish Government Health Directorates, The Scottish Government, UK

  The Cochrane Vascular editorial base is supported by the Chief Scientist Office.

Declarations of interest

AB: NIHR grant to Systematic Review Initiative at NHS Blood and Transplant. NHSBT funded a salary, training costs and provided administrative support during the review process
GO: NIHR grant to Systematic Review Initiative at NHS Blood and Transplant. NHSBT funded a salary, training costs and provided administrative support during the review process
CD: none known
LE: none known
NJW: has received research grants from the NIHR and MRC. Pfizer part‐fund a junior researcher working on a methodology project using historical data in a clinical area unrelated to this project. NJW has received honoraria from ABPI for delivering master classes on evidence synthesis. NJW has delivered a short‐course on network meta‐analysis to ICON plc, the funds from which were paid to her institution.

New