Anticoagulation and Neuraxial Anaesthesia

Transcription

1 29 May 2009 CONTENTS Anticoagulation and Neuraxial Anaesthesia G Allopi Commentator: R Naidoo Moderator: J Rubin INTRODUCTION... 3 BENEFITS OF NEURAXIAL TECHNIQUES... 4 THE COAGULATION PROCESS... 5 INCIDENCE AND AETIOLOGY OF SPINAL/EPIDURAL HAEMATOMA.. 9 THE NEED FOR CONSENSUS STATEMENTS ANTIPLATELET DRUGS Aspirin and NSAIDs ADP Receptor Antagonists GPIIb/llla Antagonists ANTICOAGULATION DRUGS Warfarin (Vitamin K Antagonists) Unfractionated Heparin (UFH) NEW ANTICOAGULANTS Low Molecular Weight Heparin (LMWH) Fondaparinux OTHER ANTICOAGULANT DRUGS Direct Thrombin Inhibitors (DTIs) Herbal Medication SPINAL EPIDURAL HAEMATOMA INFORMED CONSENT CONCLUSION REFERENCES Department of Anaesthetics Page 2 of 39

2 INTRODUCTION Recently, there have been several developments in the understanding of the coagulation process and this has lead to the development of newer anticoagulant agents in order to address the limitations of the existing drugs. The number of anticoagulant-treated patients who require elective surgery is constantly increasing. The increasing, common use of newer anticoagulant/antiplatelet agents may render the anaesthetic management of these patients more challenging and complicated. Questions have been raised about the safety of neuraxial anaesthesia in these patients, the effect of surgery and effects of discontinuation of the medication. These concerns are amplified by two potential drawbacks of the newer agents 1 : 1. Many of the newer drugs have minimal or no effect on routine coagulation tests (INR & APTT) making the detection and assessment of their anticoagulant effect difficult 2. Most of the newer drugs have no specific antidote for rapid reversal The incidence of deep vein thrombosis (DVT) as determined by postoperative venography ranges from 36% to 84% in control or placebo patients who did not receive the benefit of anticoagulation or antiplatelet therapy after hip femur or knee surgery 4. In addition, the incidence of fatal PE can run as high as 12.9% after hip fracture surgery in patients who do not receive DVT prophylaxis. Therefore, it is imperative that patients presenting these types of surgeries receive some form of perioperative anticoagulation therapy. Neuraxial anaesthesia is associated with a small but distinct risk of epidural haematoma (1 haematoma per cases); however, the risk is increased 15-fold by concomitant use of anticoagulant therapy when appropriate precautions are not taken 6. Epidural haematomas can be devastating complications, with paralysis occurring even after surgical decompression. Whatever the anticoagulant agent, the balance between the benefits (prevention of venous thrombo-embolism) and risks (major bleeding and epidural haematoma) are the central consideration. This balance not only depends on the pharmacology of each agent, but also on the dosage, timing of administration, and the type of surgery and patient characteristics. BENEFITS OF NEURAXIAL TECHNIQUES Neuraxial (spinal and epidural) anaesthesia has been demonstrated to have significant advantages compared with general anaesthesia for some surgical patients. Epidural anaesthesia confers multiple benefits including decreased intraoperative blood loss, superior postoperative analgesia, decreased surgical stress response with lower postoperative immune suppression, and decreased hypercoagulability 15. Certain surgical patients may experience as much as a 50% reduction in pulmonary embolism and 40% reduction in postoperative pulmonary infection with epidural anaesthesia 16. The total knee arthroplasty patient may obtain the additional benefit of improved rehabilitation. Some studies have shown a decreased incidence of postoperative deep vein thrombosis (DVT) formation in patients undergoing total knee arthroplasty and total hip arthroplasty with neuraxial anaesthesia when compared with patients receiving general anaesthesia 16. A recent metaanalysis by Rodgers et al. 2 involving a total of 9,559 patients and 141 studies has clearly demonstrated that outcomes are better if one uses regional rather than general anaesthesia for total hip replacement, total knee replacement, and hip fracture surgery. In this metaanalysis, there were one third fewer myocardial infarctions in the patients who received regional anaesthesia and there was a 59% reduction in the incidence of respiratory depression. Other risks, following anticoagulation use, include: increased perioperative blood loss, intrapulmonary haemorrhage following pulmonary artery catheter placement and significant bleeding following endotracheal intubation or nasogastric tube placement. Page 3 of 39 Page 4 of 39

3 THE COAGULATION PROCESS Coagulation is no longer seen as a series of separate processes, but as one integrated, time dependent process that relies on the interaction of cellular and humoral elements. Haemostasis or blood clot formation involves a series of coordinated complex interactions of injured vessels, platelets, coagulation factors, and fibrinolysis. (Figure 1) Figure 2 Platelet plug formation The primary response of haemostasis is vasospasm and activation of platelets at the site of vascular injury to form a platelet plug. Platelet activation triggers four processes: (Figure 2). adhesion (deposition of platelets on the damaged subendothelial matrix), aggregation (platelet-platelet binding), secretion (release of platelet granules), procoagulant activity (enhancement of thrombin generation) Platelet activation is mediated by a number of physiological agonists, including thrombin, collagen, ADP (adenosine diphosphate), and adrenaline. Specific receptors for these agonists exist on the platelet surface. The interaction of these receptor agonist complexes with coupling proteins triggers transmembrane signalling leading to further intracellular enzyme reactions and exposure of surface receptors. Page 5 of 39 Platelet adhesion Damage to the vascular endothelium exposes subendothelial collagen and von Willebrand s factor (vwf). Plasma vwf binding to collagen becomes a strong adhesive protein which binds or anchors circulating platelets via a platelet glycoprotein surface receptor GP Ib-IX-V to the area of vascular damage, localizing the platelet plug formation. Platelet aggregation Following activation, platelet surface protein GP IIb-IIIa undergoes a critical conformational change and becomes a high affinity receptor for fibrinogen as well as vwf and other adhesive proteins. The binding of bivalent fibrinogen molecules to activated GP IIb-IIIa on adjacent platelets bridges or links them together forming platelet aggregates. Functional activation of the IIb-IIIa receptor is the final common pathway in platelet aggregation, and its blockade is the target of the class of IIb-IIIa inhibitor antiplatelet agents. Platelet secretion After activation, platelets release a number of intracellular granules which modulate platelet interactions. These include: ADP and serotonin which will activate and recruit additional platelets; fibronectin and thrombospondin adhesive proteins that stabilize platelet aggregates; fibrinogen and vwf to promote further platelet adhesion, aggregation and fibrin clot formation; Factor V important in the coagulation process; and growth factors like Page 6 of 39

4 platelet derived growth factor that mediates tissue repair and probably atherosclerosis. In addition, Thromboxane A 2 is synthesized and released, which promotes vasoconstriction and platelet aggregation. Procoagulant activity There is a very close interaction between the coagulation cascade and activated platelets. Platelet activation results in the exposure of procoagulant anionic phospholipids (phosphatidylserine) on platelet membrane surfaces. This serves as a template to facilitate the surface assembly of coagulation factor enzyme complexes: Factor X activating complex (X-VIIIa/IXa/Ca 2+ ) and Prothrombinase complex (II Xa/Va/Ca 2+ ) in a localized catalytically efficient environment for thrombin generation (Figure 3). Thrombin generation is 300,000 times more efficient by surface complex assembly than random circulating coagulation factor interactions. The end result of these interactions is the efficient amplification and localization of the coagulation process to the area of platelet plug and vascular injury. Although it has been traditional to divide the coagulation cascade into the intrinsic and extrinsic pathways with independent trigger mechanisms, we now know that physiological coagulation is triggered by the exposure of tissue factor at the injury site. Tissue factor (TF) then immediately binds to small amounts of circulating Factor VII. This leads to rapid auto-catalytic conversion of Factor VII to additional VIIa. This TF-VIIa complex then activates Factor X directly, thereby initiating the classic extrinsic pathway and as well indirectly by activating trace amounts of Factor IX, inducing activation of the intrinsic pathway via formation of intrinsic Factor X activating complex (X-VIIIa/IXa/Ca 2+ ). This dual pathway of activation leads to more efficient generation of Factor Xa. Factor Xa can further activate additional Factor VII, amplifying the response. In essence, coagulation is triggered by tissue factor release and extrinsic pathway activation. This process is then amplified by the intrinsic pathway coagulation factors which have been activated by TF-VIIa in a positive feedback loop. Both pathways converge at the level of Factor X activation, leading to activation of the final common pathway (Figure 4). Factor Xa thus plays a central role in the coagulation process. Subsequently, Factor Xa together with Factor Va, calcium, and the platelet phospholipids form Prothrombinase complex, which activates prothrombin to thrombin, the final enzyme for the conversion of fibrinogen to fibrin. Figure 3 Page 7 of 39 Figure 4 Page 8 of 39

5 The older concept of separate intrinsic and extrinsic pathways is no longer regarded as valid. Rather, the two pathways should be seen as tightly integrated, with the tissue factor (extrinsic) pathway initiating thrombin generation and triggering the intrinsic elements to generate the thrombin burst. Thus, both components are critical if normal coagulation is to occur. INCIDENCE AND AETIOLOGY OF SPINAL/EPIDURAL HAEMATOMA In the most extensive review to date on this subject Kreppel et al. 13 analyzed 613 cases published between 1826 and In that study, the largest group (30%) was comprised of patients in whom there was no identifiable triggering factor (idiopathic or spontaneous). The second largest group included cases in anticoagulated patients (17%). Spinal epidural hematoma following spinal or epidural anaesthesia accounted for a small fraction of all cases. Only 10% of cases occurred after spinal or epidural anaesthesia. The incidence of spinal epidural hematoma is virtually impossible to pin down with accuracy but estimates have varied from 1.7% in patients suffering from spinal trauma to 1/150,000 following epidural anaesthesia to 1/1,000,000 for spontaneous (idiopathic) cases. More than 50% of the cases analyzed in this report occurred after Increased awareness, the widespread availability of modern imaging techniques such as magnetic resonance imaging and the ease of communication with electronic media may be responsible for the increasing number of cases reported in the literature. It may also be possible that increased vigilance and early surgical intervention herald better outcomes and there is, therefore, less reluctance to report such cases. In the classic review article by Vandermeulen et al. 5, they reported the occurrence of 61 spinal hematomas in patients receiving neuraxial anaesthesia between 1906 and 1994 (Table 1). At the time of anaesthetic administration, 42 of 61 (69%) of the patients developing a spinal haematoma had impaired coagulation. In 25 of the cases, some form of heparin therapy was implicated. In addition, five of the patients had undergone a major vascular procedure in which heparin was likely used, but its use was not reported on their anaesthetic record. The remaining 12 patients had a variety of conditions that could have altered their coagulation profile. Some of the conditions were thrombocytopenia, hepatic dysfunction, renal insufficiency, or the administration of another anticoagulant or plateletaltering agent. Page 9 of 39 THE NEED FOR CONSENSUS STATEMENTS Guidelines have been generated to help practitioners and protect patients but may be interpreted out of context and lead to further anxiety. In 1998, The American Society of Regional Anaesthesia (ASRA) had published its first consensus guidelines regarding regional anaesthesia in the anticoagulated patient. This was updated in 2002 and released as the ASRA 2 nd consensus statement: Regional Anaesthesia in the Anticoagulated Patient: Defining the Risks. The development of these guidelines illustrates the importance anaesthesiologists ascribe to patient advocacy and should be applauded. To quote and paraphrase the guidelines, A cook-book approach is not appropriate. Rather, decisions should be made on an individual basis, weighing the small, though definite risk of spinal hematoma with the benefits of regional anaesthesia for a specific patient. This summary statement is designed to help put the recommendations in perspective. Why is this necessary? Again, quoting from ASRA s guidelines, Practice guidelines or recommendations summarize evidence-based reviews. The rarity of spinal hematoma defies a prospective-randomized study, and there is no current laboratory model. As a result, these Consensus Statements represent the collective experience of recognized experts in the field of neuraxial anaesthesia and anticoagulation. They are based on case Page 10 of 39

6 reports, clinical series, pharmacology, haematology, and risk factors for surgical bleeding. This is an acknowledgment by the expert consensus panel that without clear understanding of the aetiology and pathogenesis of spinal epidural hematoma it is not possible to make definitive statements regarding prevention. Rather, the guidelines are an attempt to promote critical thinking and provide a basis for informed decision making in a confusing clinical landscape. ASRA is committed to reviewing and updating its guidelines regularly, as new evidence arises. ANTIPLATELET DRUGS 9 Several European groups have also formulated guidelines for the management of anticoagulated patients. See Table 2 Figure 5 Aspirin and NSAIDs Table 2 These agents interfere with coagulation by blocking the production of Thromboxane A 2 in the platelet relative to endothelial prostacyclin, resulting in decreased platelet adhesion. They also interfere with platelet-to-platelet adhesion (GPIIb/IIIa) and activation. Aspirin is widely used to reduce the risk of arterial thrombosis, especially in the presence of endothelial plaque and has found an established place in the management of myocardial ischaemia and stroke, particularly in the setting of carotid artery stenosis. Page 11 of 39 Aspirin (Acetylsalicylic acid) is the only NSAID used therapeutically for its antiplatelet action and is the most widely used antiplatelet agent in clinical practice. Its action is mediated primarily through acetylation of platelet prostaglandin synthase and is irreversible for the life of the platelet (8 10 Page 12 of 39

7 days). This prevents the platelet from producing thromboxane. As aspirin is a non-specific inhibitor of prostaglandin synthase, it also inhibits endothelial production of prostacyclin. However, the endothelial cells are able to produce new prostaglandin synthase, as we contain mitochondria, and are therefore less affected than the platelet, particularly with low-dose aspirin. Uniquely among the NSAIDs, aspirin also inhibits the GPIIb/llla receptor, thus enhancing the anti-platelet effect. Impairment of in vitro platelet function has been variably reported as lasting between 48 hours up to 10 days, but it is unclear whether or not this translates into clinical bleeding. A number of recent studies have shown no significant effect of aspirin on postoperative bleeding or the need for transfusion even in high-risk surgery. No firm conclusions can be drawn regarding the risk of surgical haemorrhage in patients taking NSAIDs who have otherwise normal coagulation function. However, it is generally accepted that the NSAIDs do not increase the risk of surgical bleeding. Some authorities suggest that aspirin -should be withdrawn for 48 hours prior to surgery but this is not a consistent recommendation. The NSAIDS also do not appear to impose any major risk to neuraxial blockade on their own, but may be additional risk factors in the presence of other anticoagulants. Where bleeding does occur in association with the use of these agents, there is some evidence that aprotinin or DDAVP may be useful. Although aspirin only has a half-life of 0.4 hours, its inhibition of platelet aggregation is irreversible and lasts for the life span of the exposed platelet. The reversal of effect requires synthesis of an adequate number of new functioning platelets (approximately 20% of total circulating platelets) and this may take 2 to 4 days. ADP Receptor Antagonists Adenosine 5-diphosphate (ADP) is a key cofactor of platelet aggregation. Three separate groups of platelet membrane-bound ADP receptors have been proposed, each with a slightly different set of activities. However, each of these receptor groups appears to be essential for platelet aggregation. The ADP-receptor antagonists permanently inhibit ADP sensitivity in platelets, enhance protein C activity and decrease the release of inflammatory markers. They have a long half-life, and increased operative bleeding has been reported in both cardiac and vascular surgical procedures, although spontaneous haemorrhage is less common with these drugs than following aspirin. Page 13 of 39 Clopidogrel (Plavix ) and Ticlopidine (Ticlic ) are the main agents in this group. They are thienopyridine derivatives that require metabolism by hepatic cytochromes before they exhibit anti-platelet activity. They permanently bind platelet ADP receptors in a non-competitive fashion, irreversibly deactivating about 70% of the ADP receptors responsible for platelet aggregation. They increase protein C activity and have antiinflammatory properties; clopidogrel also reduces formation of plateletleukocyte conjugates. The half life of these agents he is in the region of hours, but in the elderly this may be markedly extended. Thromboxane function is unaffected by these agents. Spontaneous bleeding is less common with these agents than with aspirin. GPIIb/llla Antagonists Three quite different agents are able to inhibit the GPIIb/llla receptors that are responsible for inter-platelet binding: Abciximab, Tirofiban and Eptifibatid. These intravenous drugs are currently widely used following coronary and carotid stenting procedures, particularly where drug eluting stents are used. Spontaneous bleeding is uncommon in the absence of heparin, and reports of surgical bleeding associated with these drugs are inconsistent. They are relatively short acting, and 50% recovery of platelet function occurs within 12 hours although full recovery of platelet aggregation may take up to 48 hours, with Abciximab. However, if urgent surgery is required, platelet infusions may be considered, although any remaining circulating drug will inactivate transfused platelets. Neuraxial manipulation is not recommended in the presence of these drugs. Ticlopidine displays nonlinear pharmacokinetics and reduced clearance on repeated dosing. It requires hepatic biotransformation to become active and the onset is delayed. The effect persists for the life of the platelet resulting in long duration of action. Clopidogrel undergoes extensive and rapid hydrolysis to its active metabolites. The maximum platelet inhibition of 40% to 60% occurs after 3 to 5 days; however, the onset of action can be shortened by giving a loading dose. Similar to ticlopidine, it irreversibly modifies the ADP receptor, and the platelets are affected for the remainder of their lifespan (7-10 days). Abciximab must be administered intravenously as an infusion. It has a short half-life in plasma probably related to rapid binding to the platelets. When the infusion is stopped, glycoprotein IIb-IIIa receptor occupancy on platelets decreased to 60% within the first 6 hours. Platelet function generally recovers within 24 to 48 hours; however, low levels of inhibition may last for 15 days after discontinuation of infusion. It is Page 14 of 39

8 not affected by renal or hepatic impairment. The pharmacokinetics of eptifibatide is linear at the usual dosages. It requires continuous intravenous infusion due to its short half life. Promptly after the initiation of infusion, greater than 90% of platelets are inhibited within 15 minutes. The effect on platelets is rapidly reversible such that the normal platelet function is usually restored within 4 hours. Approximately 50% of administered drug is eliminated in urine and 27% is broken down in plasma into naturally occurring amino acids. The antiplatelet effect is prolonged in patients with renal impairment. The pharmacokinetic properties of tirofiban are very similar to eptifibatide. It reaches maximum plasma concentrations and onset of action rapidly after intravenous loading infusion. It is mainly eliminated by renal excretion with limited metabolism. Thus, plasma clearance is increased by 50% in patients with severe renal impairment, however, is not affected by hepatic impairment. Platelet function is expected to return to normal in 4 to 8 hours after discontinuation of infusion. The increasing use of anti-platelet agents poses a substantial problem for anaesthetists required to handle patients who may be taking these drugs. Anti-platelet drugs are widely prescribed for a variety of conditions, notably myocardial infarction and stroke. Withdrawal of these drugs may result in significant adverse events, and is not a decision to be taken lightly since stopping anti-platelet therapy is associated with significant morbidity and mortality. With aspirin alone there does not seem to be an increase in the risk of perioperative bleeding events or in the risk of epidural haematoma. However, concomitant use of other anticoagulants, including heparin, low molecular weight heparin and other anti-platelet drugs all increase the risk of bleeding with aspirin. The key problem facing anaesthetists at present is the extensive use of anti-platelet drugs by cardiologists and vascular surgeons following the placement of intra-arterial stents. In these patients, aspirin is generally maintained for life and Clopidogrel for varying periods ranging between 6 weeks and 12 months depending on the underlying pathology and the nature of the stent used. For bare metal stents, the use of Clopidogrel for 6 weeks is generally recommended whereas for drug-eluting stents up to 12 months of Clopidogrel therapy is usually recommended. There is no clearcut indication as to when this anti-platelet drug can be safely stopped. The dilemma facing the anaesthetist is that of balancing the risks of perioperative haemorrhage against the risks of occlusive arterial events, particularly in the presence of a stent. A major risk factor in these patients Page 15 of 39 seems to be the timing of surgery after stent placement. The incidence of complications has been shown to be particularly high when surgery is performed early (<35 days) and these adverse events are mainly cardiacrelated rather than haemorrhagic, supporting the continued use of antiplatelet drugs in patients with stents. Measurement of the bleeding time before placement of a spinal or epidural anaesthetic is not indicated and it will not provide one with any useful information. Anaesthetic Management of the Patient receiving Antiplatelet Medications The following are guidelines from ASRA II (2003) 1. There is no wholly accepted test, including the bleeding time, which will guide antiplatelet therapy. Careful preoperative assessment of the patient to identify alterations of health that might contribute to bleeding is crucial. These conditions include a history of easy bruisability/excessive bleeding, female gender, and increased age. 2. NSAIDs appear to represent no added significant risk for the development of spinal hematoma in patients having epidural or spinal anaesthesia. The use of NSAIDs alone does not create a level of risk that will interfere with the performance of neuraxial blocks. 3. At this time, there do not seem to be specific concerns as to the timing of single-shot or catheter techniques in relationship to the dosing of NSAIDs, postoperative monitoring, or the timing of neuraxial catheter removal. 4. The actual risk of spinal hematoma with ticlopidine and clopidogrel and the GP IIb/IIIa antagonists is unknown. Consensus management is based on labelling precautions and the surgical, interventional cardiology/radiology experience. a. Based on labelling and surgical reviews, the suggested time interval between discontinuation of thienopyridine therapy and neuraxial blockade is 14 days for ticlopidine and 7 days for clopidogrel. b. Platelet GP IIb/IIIa inhibitors exert a profound effect on platelet aggregation. Following administration, the time to normal platelet aggregation is hours for abciximab and 4-8 hours for eptifibatide and tirofiban. Neuraxial techniques should be avoided until platelet function has recovered. GP IIb/IIIa antagonists are contraindicated within four weeks of surgery. Should one be Page 16 of 39

9 administered in the postoperative period (following a neuraxial technique), the patient should be carefully monitored neurologically. 5. The concurrent use of other medications affecting clotting mechanisms, such as oral anticoagulants, unfractionated heparin, and LMWH, may increase the risk of bleeding complications. Cyclooxygenase-2 inhibitors have minimal effect on platelet function and should be considered in patients who require anti-inflammatory therapy in the presence of anticoagulation. ANTICOAGULATION DRUGS Warfarin (Vitamin K Antagonists) Warfarin is a coumarin compound that exerts its anticoagulant effect by inhibiting the synthesis of the active forms of four Vitamin K-dependent procoagulant proteins: Prothrombin (Factor II), Factors VII, IX, and X, and two anticoagulant proteins: Protein C and Protein S. The physiological activity of these proteins is dependent on λ-carboxylation by a carboxylase enzyme, which requires reduced Vitamin K. Warfarin blocks the regeneration of reduced Vitamin K. Subsequent depletion of reduced Vitamin K results in failure to synthesize any further active forms of these coagulation proteins. An anticoagulation effect occurs within 24 hours of instituting warfarin therapy as a result of the inhibition of the production of factor VII, which has a half-life of 6 to 7 hours; but, peak anticoagulation activity is delayed for 72 to 96 hours because of the longer plasma half-lives of factors II, IX, and X Warfarin also results in the depletion of the anticoagulation proteins (protein C and protein S). Protein C also has a relatively short half-life, like factor VII. Therefore, there is a potential for the anticoagulation effects of factor VII depletion to be countered by the thrombogenic effects of reduced protein C activity during the first 24 to 48 hours of warfarin therapy. Therefore, one must keep in mind that there may actually be a thrombogenic effect to warfarin therapy during the first 24 to 48 hours of therapy. Figure 6 Reversing the effect of warfarin can be done with several modalities, depending on the urgency. In patients where rapid reversal is not required, correction of INR can be achieved with 1 to 5 mg of vitamin K given orally or intravenously. When oral vitamin K is not available, the correct amount in intravenous formulation can be administered orally. Subcutaneous administration is not recommended due to lower efficacy compared with intravenous route and unpredictable response. Normalization of the INR can be expected within 24 hours. Larger doses (ie, greater than 10 mg) of vitamin K can render the patient warfarin-resistant for up to a week or more. In acute bleeding, vitamin K should be administered along with fresh frozen plasma to rapidly reverse the anticoagulant effect. Factor concentrates may also be used (Prothrombin Complex Concentrates and recombinant factor VIIa). Ideally, the INR should be <1.5 before proceeding to invasive procedures. Page 17 of 39 Page 18 of 39

10 Abridged ASRA II Guidelines for the Regional Anaesthetic Management of the Patient Who Is Taking an Oral Anticoagulant For patients on chronic oral anticoagulation, the anticoagulant therapy must be stopped (ideally 4 5 days before the planned procedure). The PT/INR should be measured and should be allowed to reach a level of 1.5 or less before the initiation of neuraxial block. For patients receiving an initial dose of warfarin before surgery, the PT/INR should be checked before neuraxial block placement if the first dose was given more than 24 hours earlier or a second dose of oral anticoagulant has been administered. Patients receiving low-dose warfarin therapy during epidural analgesia should have their PT/INR monitored on a daily basis and checked before catheter removal if the initial dose of warfarin was administered more than 36 hours preoperatively. Neuraxial catheters should be removed when the INR is <1.5. Unfractionated Heparin (UFH) These are a heterogeneous mixture of sulphated glycosaminoglycans of varying molecular fragment lengths (average molecular weight 15,000 Da) whose major anticoagulant effect is dependent on a specific pentasaccharide sequence with a high affinity for Antithrombin III (AT), the major inhibitor of Thrombin, Factor Xa, TF-VIIa, Factor IXa, and Factor XIa. Binding of UFH to AT results in a conformational change in AT, which catalyzes by a 1000-fold, the physiologically important inactivation of Thrombin and Factor Xa. Effective Thrombin (Factor IIa) inactivation requires that the heparin fragment be long enough to bind or bridge both thrombin and AT simultaneously, an effect that occurs only if the heparin fragment chain exceeds 18 monosaccharide units, commonly found in UFH. In contrast, this type of simultaneous bridging is not required for AT inactivation of Factor Xa. The dose requirement of UFH is also unpredictable due to variable binding to plasma lipoprotein, globulins, and fibrinogen. The onset is immediate following intravenous infusion, but slightly delayed when given subcutaneously. The elimination involves a combination of a rapidly saturable mechanism at low doses and a slower first-order mechanism at higher doses. As a result, the half-life increases as the dose increases. The effect may be prolonged with renal and hepatic impairment. Page 19 of 39 Protamine sulphate is a highly basic protein which combines with strongly acidic heparins and rapidly neutralizes both the anti-xa and the anti-iia anticoagulant effect of UFH. Since plasma UFH concentrations decrease rapidly, the dose of protamine required also decreases as time elapses. If UFH is given as an infusion, the dose of protamine required is based on the amount administered in the previous 4 hours at 1 mg per 100 units of UFH. If 90 minutes (ie, one half-life) have elapsed since the last bolus injection, then 0.5 mg per 100 units of UFH should be given. Note that the half-life of protamine is shorter than UFH, when UFH is given subcutaneously, thus repeated doses of protamine may be required. Protamine should only be given by slow IV infusion, with no more than 50 mg over a 10-minute period. Protamine sulphate can only partially (60%) neutralize the anticoagulant effect (anti-xa) of commercially available LMWHs because of the presence of varying amounts of low-sulphated LMWH molecules which are protamine resistant. Abridged ASRA Guidelines for Use of Neuraxial Techniques in Patients Receiving Low-dose Subcutaneous Unfractionated Heparin During subcutaneous minidose prophylaxis (5,000 units 2 hours before surgery), there is no contraindication to the use of neuraxial techniques. The risk of neuraxial bleeding may also be reduced by delaying the heparin injection until 1 to 2 hours after placement of the neuraxial block. There may be an increased risk of neuraxial bleeding in debilitated patients or in patients who have received prolonged UH therapy. Because HITT may occur during heparin administration, patients receiving heparin for greater than 4 days should have a platelet count assessed before neuraxial block. Avoid neuraxial techniques in patients with other coagulopathies. Heparin administration should be delayed for 1 hour after needle placement. Remove the catheter 1 hour before any subsequent heparin administration or 2 to 4 hours after the last heparin dose. Monitor the patient postoperatively to provide early detection of motor blockade and consider the use of minimal concentrations of local anaesthetics to facilitate the early detection of a spinal hematoma. Although the occurrence of a bloody or difficult neuraxial needle placement may increase risk, there are no data to support mandatory cancellation of a case. Clinical judgment is needed. If a decision is made to proceed, full discussion with the surgeon and careful postoperative monitoring is warranted. Page 20 of 39

11 The European Guidelines for Patients Who Are Receiving or Will Receive Unfractionated Heparin and a Neuraxial Block The only substantial difference between the European and ASRA guidelines is the fact that the Europeans suggest that one wait 4 hours versus 2 hours after the subcutaneous injection of 5,000 units of UH before placing a neuraxial block. The Safety of Neuraxial Anaesthesia in the Patient Receiving Therapeutic and Full Anticoagulation With Unfractionated Heparin (20,000 to 30,000 Units Intravenously) In 1998, Sanchez and Nygard reported 558 cardiac surgery patients who had epidural catheters placed following strict guidelines. These included placement of the epidural catheters the day before the surgery and limiting attempts at catheter placement to two attempts. There was a zero incidence of spinal hematoma formation in this study. Abridged ASRA Guidelines for the Administration of Neuraxial Anaesthesia in the Patient Fully Anticoagulated with Unfractionated Heparin Currently, insufficient data and experience are available to determine if the risk of neuraxial hematoma is increased when combining neuraxial techniques with the full anticoagulation effects required during cardiac surgery. Prolonged therapeutic anticoagulation appears to increase the risk of spinal hematoma formation, especially if combined with other anticoagulants or thrombolytics. Therefore, neuraxial blocks should be avoided in this clinical setting. If systemic anticoagulation therapy is begun with an epidural catheter in place, it is recommended that one delay catheter removal for 2 to 4 hours after therapy discontinuation and evaluation of coagulation status. NEW ANTICOAGULANTS Low Molecular Weight Heparin (LMWH) Low molecular weight heparins are derived from unfractionated heparin by chemical or enzymatic depolymerisation to yield fragments about one-third the size of UFH. These smaller heparin fragments ( Da) still contain the crucial AT-binding pentasaccharide sequence and are able to effectively catalyze the inactivation of Factor Xa. However, their smaller fragment chain length preclude LMWHs from the simultaneous bridging of AT and Thrombin necessary for Thrombin (IIa) inactivation. Therefore, most LMWHs have high anti-xa activity and minimal anti-iia activity (anti-xa/anti- IIa ratio ~2-4:1). Thus, the activated partial thromboplastin time (aptt), a measure of antithrombin (antifactor IIa) activity, is not useful in measuring the activity of LMWHs, which requires instead a specific anti-xa assay. LMWHs have predictable pharmacokinetics such that they can be given without close monitoring of anticoagulant effect. Several pharmacologically distinct agents are commercially available (enoxaparin, dalteparin, ardeparin, nadroparin, tinzaparin and reviparin). However, their anti-iia and anti-xa activities are sufficiently different that they are not considered interchangeable. The onset is typically slower and duration of action is longer than UFH, requiring once to twice daily dosing. They are eliminated renally, thus the half-life is expected to be prolonged in patients with renal insufficiency. The LMWHs cannot be neutralized by protamine. Anaesthetic Management of the Patient Receiving Low Molecular Weight Heparin (LMWH) Anesthesiologists in North America can draw on the extensive European experience to develop practice guidelines for the management of patients undergoing spinal and epidural blocks while receiving perioperative LMWH. All consensus statements contained herein respect the labelled dosing regimens of LMWH as established by the FDA. Although it is impossible to devise recommendations that will completely eliminate the risk of spinal hematoma, previous consensus recommendations have appeared to Page 21 of 39 Page 22 of 39

12 improve outcome. Concern remains for higher dose applications, where sustained therapeutic levels of anticoagulation are present. 1. Monitoring of the anti-xa level is not recommended. The anti-xa level is not predictive of the risk of bleeding and is, therefore, not helpful in the management of patients undergoing neuraxial blocks. 2. Antiplatelet or oral anticoagulant medications administered in combination with LMWH may increase the risk of spinal hematoma. Concomitant administration of medications affecting haemostasis, such as antiplatelet drugs, standard heparin, or dextran represents an additional risk of hemorrhagic complications perioperatively, including spinal hematoma. Education of the entire patient care team is necessary to avoid potentiation of the anticoagulant effects. 3. The presence of blood during needle and catheter placement does not necessitate postponement of surgery. However, initiation of LMWH therapy in this setting should be delayed for 24 hours postoperatively. Traumatic needle or catheter placement may signify an increased risk of spinal hematoma, and it is recommended that this consideration be discussed with the surgeon. 4. Preoperative LMWH a. Patients on preoperative LMWH thromboprophylaxis can be assumed to have altered coagulation. In these patients needle placement should occur at least hours after the LMWH dose. b. Patients receiving higher (treatment) doses of LMWH, such as enoxaparin 1 mg/kg every 12 hours, enoxaparin 1.5 mg/kg daily, dalteparin 120 U/kg every 12 hours, dalteparin 200 U/kg daily, or tinzaparin 175 U/kg daily will require delays of at least 24 hours to assure normal haemostasis at the time of needle insertion. c. Neuraxial techniques should be avoided in patients administered a dose of LMWH two hours preoperatively (general surgery patients), because needle placement would occur during peak anticoagulant activity. 5. Postoperative LMWH Patients with postoperative initiation of LMWH thromboprophylaxis may safely undergo single-injection and continuous catheter techniques. Management is based on total daily dose, timing of the first postoperative dose and dosing schedule. (a) Twice daily dosing. This dosage regimen may be associated with an increased risk of spinal hematoma. The first dose of LMWH should be administered no earlier than 24 hours postoperatively, regardless of anaesthetic technique, and only in the presence of adequate (surgical) haemostasis. Indwelling Page 23 of 39 (b) Fondaparinux catheters should be removed prior to initiation of LMWH thromboprophylaxis. If a continuous technique is selected, the epidural catheter may be left indwelling overnight and removed the following day, with the first dose of LMWH administered at least two hours after catheter removal. Single daily dosing. This dosing regimen approximates the European application. The first postoperative LMWH dose should be administered 6-8 hours postoperatively. The second postoperative dose should occur no sooner than 24 hours after the first dose. Indwelling neuraxial catheters may be safely maintained. However, the catheter should be removed a minimum of hours after the last dose of LMWH. Subsequent LMWH dosing should occur a minimum of 2 hours after catheter removal. This is one of the new classes of selective Factor Xa indirect inhibitors recently approved for clinical use as antithrombotic prophylaxis following orthopaedic surgery. It is also increasingly being used for treatment of deep vein thrombosis and pulmonary embolism and as alternative antithrombotic management in Heparin-induced Thrombocytopenia (HIT). Fondaparinux 2.5 mg per day reduces the incidence of DVT in patients undergoing total knee arthroplasty by 55.2% compared with patients receiving enoxaprin 30 mg twice each day for venous thrombo embolus prophylaxis. It is a totally synthetic analogue of the pentasaccharide sequence that binds to AT with high affinity, catalyzing (300-fold) the inactivation of Factor Xa, and thereby reducing thrombin generation. However, its short chain length renders it incapable of inactivating thrombin (Factor IIa) directly. Fondaparinux pharmacokinetics is linear and predictable. It specifically binds to AT and does not bind significantly to other plasma proteins or red blood cells. The onset is rapid after subcutaneous injections, and the half-life is sufficiently long for once daily administration. It is primarily eliminated via the kidneys with up to 77% of the dose recovered in urine as unchanged drug. Total clearance is 25% lower with mild renal impairment, 40% lower with moderate impairment, and 55% with severe impairment Page 24 of 39

13 Anaesthetic management of the patient on Fondaparinux The FDA released fondaparinux with a black box warning similar to that of the LMWHs and heparinoids. The actual risk of spinal hematoma with fondaparinux is unknown. Consensus statements are based on the sustained and irreversible antithrombotic effect, early postoperative dosing, and the spinal hematoma reported during initial clinical trials. Close monitoring of the surgical literature for risk factors associated with surgical bleeding may be helpful in risk assessment and patient management. ASRA II consensus statement (2003) Until further clinical experience is available, performance of neuraxial techniques should occur under conditions utilized in clinical trials (single needle pass, atraumatic needle placement, avoidance of indwelling neuraxial catheters). If this is not feasible, an alternate method of prophylaxis should be considered. ASA 2005 suggested guidelines Regional anaesthesia may be safely used as the surgical anaesthetic when fondaparinux is administered in a dose of 2.5 mg subcutaneously, and the first dose is administered 8 or more hours after the completion of surgery The one case of spinal hematoma formation reported in the literature 22 involved a patient who had had an attempted epidural catheter placement, which involved five or more needle passes and the subsequent administration of 6 mg of fondaparinux after the completion of surgery (the patient was a subject in a dose response trail). This was more than twice the recommended dose of fondaparinux. Although the surgery was ultimately performed under general anaesthesia, this case report would suggest that epidural anaesthesia may carry more risk than spinal anaesthesia performed with a fine-gauge needle (25 29 g). If epidural anaesthesia is administered, it is imperative that the catheter be removed immediately after the completion of surgery and catheters should not be retained for the administration of postoperative analgesia. Finally, if one should experience difficulty in placing a spinal or epidural block, more than two needle passes, or if the lumbar puncture was bloody, one must communicate these facts with one s surgical colleague. Discuss with the surgeon the risks and benefits of using fondaparinux as the antithrombotic in such cases and suggest alternatives. Although warfarin may be less efficacious, it may be the safer alternative after a traumatic lumbar puncture. Page 25 of 39 The optimum time to the administration of the first dose of fondaparinux has been determined to be between 6 and 8 hours after surgery. This time interval to the first dose was based on the outcomes from five studies However, authors suggest that clinicians use the upper recommended time limit and wait 8 hours after the completion of surgery before administering the first dose of fondaparinux. One spinal hematoma has already been reported using this time interval in a patient receiving twice the recommended dose of fondaparinux. As previously mentioned, this fondaparinux related hematoma occurred in a patient who had had an attempted epidural catheter placement. Fonadaparinux is contraindicated until surgical haemostasis is achieved, but can be administered 6 hours postoperatively and is typically repeated daily for 5 to 11 days but has been studied for up to 1 month postoperatively 16. The manufacturer does not recommend using fonadaparinux before neuraxial anaesthesia or with an indwelling epidural catheter in place. Newer studies New articles and editorials 23 have shown increased use of fondaparinux, including the use of epidural catheters for several days postoperatively. Most articles do not advocate the use of epidural catheters postoperatively. Neuraxial blocks have been successfully performed hours after the last dose. Fondaparinux has been restarted 8 hrs following surgery and removal of indwelling catheters. Clinical experience with the combination of fonadaparinux and neuraxial anaesthesia is growing. SAFETY OF FONDAPARINUX IN PATIENTS UNDERGOING MAJOR ORTHOPAEDIC SURGERY WITH A NEURAXIAL CATHETER FOR ANAESTHESIA/ANALGAESIA: THE EXPERT OBSERVATIONAL STUDY N. Rosencher, Florence, Italy: May 2007 Continuous neuraxial or deep peripheral nerve blockade used to provide postoperative analgesia after major orthopaedic surgery is associated with a risk of spinal or perineural haematoma, especially in patients concomitantly receiving anticoagulants. Limited data on the use of fondaparinux in surgical patients in whom this procedure is performed are available. The EXPERT trial was an observational international study in patients undergoing major orthopaedic surgery designed to evaluate the overall Page 26 of 39

14 efficacy and safety of once-daily 2.5 mg fondaparinux initiated 6 to 12 hours post-operatively and administered for 4±1 weeks after surgery. A 48-hour "therapeutic window" was applied in patients in whom a neuraxial/deep peripheral indwelling catheter was placed: one of the planned doses of fondaparinux was omitted, the catheter was removed 36 hours after the previous fondaparinux dose, and the next fondaparinux dose administered 12 hours after catheter removal. The primary endpoints were symptomatic venous thromboembolism (VENOUS THROMBOEMBOLISM) and major bleeding 5±1 weeks after surgery. Overall, 5704 patients (mean age ± SD: 66 ± 12 years) were recruited between July 2003 and October They underwent surgery for total hip replacement (52%, n=2941), knee replacement (40%, n=2263), hip fracture (6%, n=353), or other indications (3%, n=148). Fondaparinux was given for a median of 35 (range: 1 105) days. Many operations (62%) were performed under regional anaesthesia only. A neuraxial or deep peripheral nerve block catheter was placed in 29% (n=1630) of patients. It was removed between one and two days after surgery in 43% (706/1626), and between three and six days after surgery in 57% (920/1626). Overall, the rate of symptomatic venous thromboembolism was 1.0% (54/5387); it was 0.8% (13/1535) in patients with catheter and 1.1% (41/3852) in patients without catheter, giving an odds ratio of 0.79 (95% CI: 0.42 to 1.49) in favour of patients with a catheter. The rate of major bleeding was 0.8% (42/5382) overall, 0.5% (7/1532) in patients with catheter and 0.9% (35/3850) in patients without catheter. No spinal or perineural hematomas or nerve damage were reported. At 5±1 weeks, 23 (0.4%) patients had died. In conclusion, 2.5 mg fondaparinux given daily for 4±1 weeks after major orthopaedic surgery was both effective and safe in routine practice. OTHER ANTICOAGULANT DRUGS Direct Thrombin Inhibitors (DTIs) Thrombin inhibitors bind directly to the active site of thrombin and block interaction with its substrate, fibrinogen, independent of AT. Unlike heparin and LMWH, DTIs can inactivate clot-bound thrombin which promotes further thrombus growth as well as circulating thrombin. Hirudin is a naturally occurring anticoagulant isolated from leech salivary gland secretions. Desirudin (Revasc, Aventis Pharmaceuticals) lepirudin (Refludan, Aventis Pharmaceuticals) and bavilirudin (Angiomax) are new recombinant hirudin derivatives. These compounds function as direct thrombin inhibitors, blocking both its catalytic and the substrate binding sites. Hirudin has several advantages over heparin, including action independent of antithrombin III, better predictability of dose response, minimal effect on platelets and no risk for HIT. In European trials, desirudin has been shown to decrease the incidence DVT after THA by 36% when compared with single day enoxaprin dosing, but this benefit may not be seen in the United States where enoxaprin typically dosed using a twice daily regimen. Argatroban is a synthetic direct thrombin inhibitor with activity similar to hirudin derivatives. It is currently approved by the USFDA for use after percutaneous coronary intervention for patients with a history of HIT, but has been investigated for use for DVT prophylaxis. It shares a similar disadvantage to hirudin derivatives that no antidote or reversal agent exists, but because of its 24-minute half-life, this may be less clinically significant for agratroban. Although potentially efficacious for venous thromboembolism prophylaxis these medications do not have USFDA approval for this indication at this time. There is insufficient experience with these agents to make recommendations regarding their use in patients undergoing neuraxial anaesthesia. The Second ASRA Consensus Conference on Neuraxial Anaesthesia and Anticoagulation was unable to propose guidelines and caution in using these agents in patients receiving neuraxial anaesthesia is prudent. Page 27 of 39 Page 28 of 39

15 Herbal Medication There is a widespread use of herbal medications in surgical patients. Morbidity and mortality associated with herbal use may be more likely in the perioperative period because of the polypharmacy and physiological alterations that occur. Such complications include bleeding from garlic, ginkgo and ginseng, and potential interaction between ginseng-warfarin. Because the current regulatory mechanism for commercial herbal preparations sold in the United States does not adequately protect against unpredictable or undesirable pharmacological effects, it is especially important for anaesthesiologists to be familiar with related literature on herbal medications when caring for patients in the perioperative period. Garlic Garlic inhibits in vivo platelet aggregation in a dose-dependent fashion. The effect of one of its constituents, ajoene, appears to be irreversible and may potentiate the effect of other platelet inhibitors such as prostacyclin, forskolin, indomethacin and dipyridamole. Although these effects have not been consistently demonstrated in volunteers, there is one case in the literature of an octogenarian who developed a spontaneous epidural hematoma that was attributed to heavy garlic use (Rose, 1990). Ginkgo Ginkgo is derived from the leaf of Ginkgo biloba. Ginkgo inhibits plateletactivating factor. Clinical trials in a small number of patients have not demonstrated bleeding complications, but four reported cases of spontaneous intracranial bleeding and one case of spontaneous hyphema have been associated with ginkgo use. Based upon the pharmacokinetic data, normalization of coagulation should occur 36 hours after discontinuation of ginkgo (Chung, 1987). Ginseng There is a concern of ginseng s effect on coagulation pathways. Ginsenosides inhibit platelet aggregation in vitro and prolong both thrombin time and activated partial thromboplastin time in rats. These findings await confirmation in humans. Although ginseng may inhibit the coagulation cascade, ginseng use was associated with a significant decrease in warfarin anticoagulation in one reported case (Janetzky, 1997). The pharmacokinetics of ginsenosides suggest that 24 hours is required to allow resolution of ginseng s effect on haemostasis. Page 29 of 39 Anaesthetic management of the patient receiving herbal therapy Herbal drugs, by themselves, appear to represent no added significant risk for the development of spinal hematoma in patients having epidural or spinal anaesthesia. This is an important observation since it is likely that a significant number of our surgical patients utilize alternative medications preoperatively and perhaps during their postoperative course. There is no wholly accepted test to assess adequacy of haemostasis in the patient reporting preoperative herbal medications. Careful preoperative assessment of the patient to identify alterations of health that might contribute to bleeding is crucial. Data on the combination of herbal therapy with other forms of anticoagulation are lacking. However, the concurrent use of other medications affecting clotting mechanisms, such as oral anticoagulants or heparin, may increase the risk of bleeding complications in these patients. SPINAL EPIDURAL HAEMATOMA Although an association between anticoagulation (or bleeding diatheses) and spinal epidural hematoma exists (and is the premise behind ASRA s guidelines), it is not clear what significance it should be given. Kreppel et al. 13 state, The aetiology of spinal haemorrhage with subsequent spinal compression is unclear in many cases. All of the etiological factors mentioned are very common in the general population and cannot be the sole cause of spinal hematoma ( in any individual case ), considering their high overall frequency. Studies suggest, that several of these factors must interact in order to cause spinal haematoma and that all the etiological factors proposed to date merely reflect the need to find an explanation and are not likely the true cause of spinal haematoma. Risk factors associated with spinal epidural haematoma formation are 26 : Female gender Advanced age Orthopaedic patients Ankylosing spondylitis Renal insufficiency Coagulopathy or use of antithrombotic drugs Multiple needle passes Catheter insertion and withdrawal Page 30 of 39

16 Different pathogenetic mechanisms have been proposed for the development of spinal epidural hematoma. These include epidural venous bleeding, epidural arterial bleeding, and bleeding from vascular malformations. Each theory has its proponents and its detractors. Regardless of which theory is favoured, anticoagulation may magnify the risk. It is important, therefore, to make a distinction between an association and cause and effect. Proponents of thoracic epidural analgesia for patients having coronary artery bypass surgery may prefer to implicate an arterial origin. The rarity of spinal epidural hematoma may then be explained by the rarity of inadvertently damaging an artery. Therefore, strict adherence to thromboprophylaxis guidelines may not be necessary or protective. Should arterial injury be involved, advances in ultrasound technology may allow an opportunity to avoid injury by identifying a needle path devoid of vascular structures. Proponents of venous bleeding as the cause of spinal epidural hematoma have difficulty explaining the disparity or poor correlation between the incidence of bleeding during neuraxial regional anaesthesia and that of compressive spinal epidural hematoma. Venous bleeding occurs frequently and cerebrospinal fluid pressure should effectively tamponade bleeding from a low pressure source. Major complications of central neuraxial block: report on the Third National Audit Project of the Royal College of Anaesthetists. T. M. Cook, British Journal of Anaesthesia 2009; 102 (2): This is the most current review of the complications associated with central neuraxial blocks. It was estimated that central neuraxial blocks were performed, nationally, over 12 months. During this time 52 major complications were included in the study. Importantly, there were 5 reports of spinal haematomas, 4 of them required decompression. This study, however, did not document the use of anticoagulants, the difficulty of the needle insertion or the level of experience of the person performing the anaesthetic. All the people who developed haematomas were elderly, with multiple co-morbidities and undergoing major surgery. This study gave an incidence of 1: for the development of spinal haematomas. Management of spinal epidural haematoma Vandermeulen and colleagues indicate that patients who are developing a neuraxial haematoma initially complain of new onset of numbness, weakness, or radicular back pain, and they ascertained that muscle weakness was the first neurologic symptom in 46% of patients with spinal hematoma and that sensory deficit was the presenting symptom in 14% of the patients. Prompt recognition and treatment of this condition is essential to optimize recovery of neurologic function in these patients. An immediate magnetic resonance image should be obtained in every patient who develops newonset neurologic deficits after the placement of a neuraxial block or removal of an epidural catheter. If the magnetic resonance image identifies the presence of a neuraxial hematoma, immediate surgical decompression is the treatment of choice. Vandermeulen et al. reported the results of 13 patients who had a decompressive laminectomy to evacuate their spinal hematoma within 8 hours of the development of paraplegia. Good or partial recovery of neurologic function was obtained in 77% of these patients (10 of 13). However, if the surgery was delayed for more than 24 hours, only 15% (two of 12) obtained good recovery, and the majority of these unfortunate patients never regained any neurologic function. Page 31 of 39 Page 32 of 39

17 INFORMED CONSENT How do we confront such controversies in everyday practice? How can we balance the risks and benefits responsibly? The primary reason such a rare complication generates intense debate relates to its tragic consequences, death or paralysis. Patients are entitled to know the existence and precise nature of any controversy that may affect their care. Furthermore, this information should be put into perspective and the proper context for them. This is the art and science of informed consent. The association of anticoagulation with epidural hematoma formation cannot help settle the question of site of origin of bleeding. It will exist regardless of whether the origin of the bleeding is venous or arterial. Without such basic knowledge regarding pathogenesis definitive strategies for prevention and management cannot be made. Like anticoagulation, trauma and technical difficulty have been associated with the occurrence of spinal epidural hematoma. However, unlike anticoagulation; few recommendations have been formulated regarding this co-factor. Why? Do we remain unconvinced of the association? Is technical difficulty equivalent to trauma? It is possible that even a single needle pass may injure a vein or artery and are we, therefore, convinced it is a problem without remedy? Page 33 of 39 Our profession, through its professional bodies, needs to direct more attention to performing this well. Most often, proper informed consent is only missed when a complication arises (lack of good consent is never missed!). Most anaesthesiologists will never personally experience the catastrophic consequences to their patient and to themselves and accordingly they may find it difficult to empathize with those unfortunate few who have. This may lead to a cavalier attitude and approach to our obligation to inform patients regarding rare but devastating complications and makes it difficult to improve the process. Where controversy and doubt exist we should learn to more effectively recruit patients to share responsibility for difficult decisions affecting their care and welfare. This is a critical element of individualized decisions formulated after an evidence-based assessment of risk and benefit. Clearly, the best-case scenario would generate unequivocal recommendations based upon valid information regarding aetiology and pathogenesis. As the ASRA guidelines suggest, this is not likely going to occur any time soon. The aetiology and pathogenesis of spinal epidural hematoma remain a mystery. Nonetheless, there may be some things worth doing right now. As suggested by Kreppel et al an international registry should be created allowing timely, comprehensive study and analysis. This type of registry would acknowledge the difficulty in studying such rare events and should facilitate rapid dissemination of relevant information. Legislation is desperately required to facilitate this process. Animal models should be developed and clinical investigations should be initiated to study the likely pathogenetic factors implicated in spinal epidural hematoma formation and progression and their relative importance. Page 34 of 39

18 Investigation in these areas should be encouraged by our professional organizations. National and international professional organizations (medical and legal) should assist practitioners to develop best practices for obtaining informed consent. At the moment there is little consensus on what patients should be told or on how patients can be recruited to help make difficult decisions. An interdisciplinary approach should be taken regarding all issues (e.g., neurology, neurosurgery, anaesthesia, medical ethics) to widen the net, to focus our resolve, and to enhance our understanding regarding diagnosis and treatment. Our current state of confusion demands this. For example, prompt diagnosis and surgical intervention seem to offer the best hope of a promising outcome. However, successful conservative management of spontaneous spinal epidural hematoma and spinal epidural hematoma following epidural analgesia has been described recently. To further confuse the issue of management and pathogenesis, full spontaneous recovery from flaccid paraplegia in a patient with compressive spinal epidural hematoma following thoracic epidural catheterization, an extremely unlikely scenario, has been described. CONCLUSION Vigilance must not be restricted to patients that have had spinal or epidural anaesthesia and to those on anticoagulants. This implies physicians from all disciplines must be vigilant. Without a multidisciplinary effort by all physicians patients will continue to slip through our diagnostic net and suffer as a result. An attempt should be made to define what is meant by trauma or technical difficulty. If we believe technical difficulty contributes to the risk of spinal epidural hematoma, we should establish technical standards and promote research into methods of facilitating easier and safer neuraxial anaesthesia. The true incidence, aetiology, and pathogenesis of spinal epidural hematoma, along with optimal preventative and management strategies remain elusive. Until more information is available the entire medical profession must remain vigilant and informed. We must stay on alert for high-risk clinical scenarios (new anticoagulants, multiple concurrent risk factors). However, we must not lose sight of the possibility that we may have been barking up the wrong tree: And if the blind lead the blind, both shall fall into the ditch. Page 35 of 39 Page 36 of 39

19 REFERENCES 1. Jay RM, Lui P. How anticoagulants work. Techniques in regional anaesthesia and pain management 2006; 10: Rodgers A, Walker N, Schug S, et al.: Reduction of post-operative mortality and morbidity with epidural and spinal anaesthesia: Results from overview of randomized trials. BMJ 2000; 321: Broadman LM. Anticoagulation and regional anaesthesia. American Society of Anaesthesiologists Guidelines 2005; 33: Geerts WH. Prevention of venous thromboembolism. Chest 2001; 119: Vandermeulen EP, Van Aken H, Vermylen J. Anticoagulants and spinal epidural anaesthesia. Anaesthesia Analg 1994; 79: Rosencher N. Selected new antithrombotic agents and neuraxial anaesthesia for major orthopaedic surgery: management strategies. Anaesthesia 2007; 62: Vitin AA. Anaesthetic implications of the new anticoagulant and antiplatelet drugs. Journal of clinical anaesthesia 2008; 20: Neuraxial anaesthesia and anticoagulation 2 nd consensus conference on neuraxial anaesthesia and anticoagulation. American Society of Regional Anaesthesia James MFM. Platelet antagonists. Part II Anaesthetics Refresher Course I.A Jobert. The Pharmacology of coagulation. Part I Anaesthetics Refresher Course Kopp SL. Anticoagulation in pregnancy and neuraxial blocks. Anaesthesiology Clinics 2008; 26: James MFM. Outcome following neuraxial blockade. Part II Anaesthetics Refresher Course Kreppel D: Spinal hematoma: a literature survey with meta-analysis of 613 patients. Neurosurgery Rev 2003; 26: Lang SA. Spinal epidural haematoma: Still an enigma (Editorial). Journal of Clinical Anaesthesia 2004; 16: Liu S. Epidural anaesthesia and analgesia: Their role in postoperative outcome. Anaesthesiology 1995; 82: Heitz JW. Neuraxial Anaesthesia and Anticoagulants. Techniques in Orthopaedics 2008; 23: Lassen MR, Bauer KA, Eriksson BI, Turpie AGG, for the European Pentasaccharide Hip Elective Surgery Study (EPHESUS) Steering Committee: Postoperative fondaparinux versus preoperative Page 37 of 39 enoxaparin for prevention of venous thromboembolism in elective hipreplacement surgery: A randomized double-blind comparison. Lancet 2002; 359: Turpie AGG, Bauer KA, Eriksson BI, Lassen MR, for the PENTATHLON 2000 Study Steering Committee: Postoperative fondaparinux versus postoperative enoxaparin for prevention of venous thromboembolism in elective hip-replacement surgery: A randomized doubleblind comparison. Lancet 2002; 359: Turpie AGG, Gallus AS, Hoek JA, for the Pentassacharide Investigators: A synthetic pentassacharide for the prevention of deepvein thrombosis after total hip replacement. N Engl J Med 2001; 344: Eriksson BI, Bauer KA, Lassen MR, Turpie AGG, for the Steering Committee of the Pentassacharide in Hip-Fracture Surgery Study: Fondaparinux compared with enoxaparin for the prevention of venous thromboembolism after hip-fracture surgery. N Engl J Med 2001; 345: Horlocker TT. Analgesia without paraplegia: Neuraxial anaesthesia and anticoagulation. CONFERENCIAS MAGISTRALES 2006; 29: Hull R, Pineo G: A synthetic pentasaccharide for the prevention of deep-vein thrombosis. NEngl J Med 2001; 345: Jacques. Do We Really Need an Interval Between Administering Fondaparinux and Removing a Lumbar Plexus Catheter? Anaesthesia and analgesia 2009; 108: Singelyn F, Felicissimo P, Piovella F, Van Aken HK, Rosencher N. Use of neuraxial catheter for anaesthesia analgesia in major orthopedic lower limb surgery Patients treated with prophylactic dose of fondaparinux in routine practice (Abstract). Journal of Thrombosis and Haemostasis 2005; 3: Cook TM. Major complications of central neuraxial block: report on the Third National Audit Project of the Royal College of Anaesthetists. British Journal of Anaesthesia 2009; 102 (2): Gogarten W. The influence of new antithrombotic drugs on regional anaesthesia. Current Opinion Anaesthesiology 2006; 19: Page 38 of 39

20 Page 39 of 39