AUTHOR: Honorio T. Benzon

Intraspinal Hematoma

The incidence of intraspinal hematoma is approximately 0.1 per 100,000 patients per year.(1) It is more likely to occur in anticoagulated or thrombocytopenic patients, patients with neoplastic disease, or in those with liver disease or alcoholism.(2) The incidence of neurologic dysfunction resulting from hemorrhagic complications associated with neuraxial blockade is estimated to be

Antiplatelet Therapy

Antiplatelet medications inhibit platelet cyclo-oxygenase and prevent the synthesis of thromboxane A2. Thromboxane A2 is a potent vasoconstrictor and facilitates secondary platelet aggregation and release reactions. An adequate, although potentially fragile, clot may form.(5) Platelet function in patients receiving antiplatelet medications should be assumed to be decreased for 1 week after treatment with aspirin and 1 to 3 days with nonsteroidal anti-inflammatory drugs (NSAIDs). New platelets are produced every day, and these new platelets partly explain the relative safety of per- forming neuraxial procedures in these patients.

Although Vandermeulen et al (6) implicated antiplatelet therapy in 3 of the 61 cases of spinal hematoma occurring after spinal or epidural anesthesia, the results of several large studies demonstrated the relative safety of neuraxial blockade in combination with antiplatelet therapy. The Collaborative Low-Dose Aspirin Study in Pregnancy Group(7) included 1422 high-risk obstetric patients who were administered 60 mg aspirin daily and underwent epidural anesthesia without any neurologic sequelae. The studies of Horlocker et al,(8,9) of approximately 1000 patients in each study, showed no spinal hematomas, although blood was noted during needle or catheter placement in 22% of the patients. A later study in patients who were on NSAIDs and underwent epidural steroid injections did not develop the signs and symptoms of intraspinal hematoma.(10) A review of the case reports of intraspinal hematoma in patients on aspirin and NSAIDs showed complicating factors that in luded concomitant heparin administration, epidural venous angioma, and technical difficulty when performing the procedure.(11)

The thienopyridine drugs, ticlopidine and clopidogrel, prevent platelet aggregation by inhibiting adenosine diphosphate (ADP) receptor-mediated platelet activation. Ticlopidine is rarely used because it causes neutropenia, thrombocytopenic purpura, and hypercholesterolemia. Clopidogrel is preferred because of its increased safety profile and proven efficacy. The maximal inhibition of ADP-induced platelet aggregation with clopidogrel occurs 3 to 5 days after the initiation of a standard dose (75 mg), but within 4 to 6 hours after the administration of a large loading dose of 300 to 600 mg.(12) There is a case report of spinal hematoma in a patient on ticlopidine(13) and a case of quadriplegia in a patient on clopidogrel, diclofenac, and possibly aspirin.(14)

Neuraxial blocks can be safely performed on patients taking aspirin or NSAIDs.(4) It is safe to perform neuraxial blocks on patients taking cyclo-oxygenase (COX)-2 inhibitors. For the thienopyridine drugs, it is recommended that clopidogrel be discontinued for 7 days and ticlopidine for 10 to 14 days before a neuraxial injection. It is possible for epidural catheters to be removed or neuraxial injections to be performed 5 days, and not 7 days, after clopidogrel is discontinued.(15) If a neuraxial injection is to be performed in a patient on clopidogrel before 7 days of discontinuation, a P2Y12 assay, a new assay of residual antiplatelet activity, can be performed;

Here is a summary of current recommendations:

1. Neuraxial blocks can be performed on patients taking aspirin or NSAIDs.(4)

2. It is safe to perform neuraxial blocks on patients taking COX-2 inhibitors.

3. For the thienopyridine drugs, the ASRA recommendation is that clopidogrel be discontinued for 7 days and     ticlopidine for 10 to 14 days before a neuraxial injection.

4. Epidural catheters can be removed safely and neuraxial injections can be performed 5 days (not 7 days, as once     advised) after clopidogrel is discontinued.(15)

5. If a neuraxial injection is to be performed in a patient on clopidogrel before 7 days of discontinuation, a P2Y12    assay, a new assay of residual antiplatelet activity, is performed;

Oral Anticoagulants

Warfarin exerts its anticoagulant effect by interfering with the synthesis of the vitamin K-dependent clotting factors (VII, IX, X, and thrombin).17 It also inhibits the anticoagulants protein C and S. Factor VII and protein C have short half-lives (6-8 hours), and the prolongation of the international normalized ratio (INR) during the early phase of warfarin therapy is the result of the competing effects of reduced factor VII and protein C.(18) Adequate anticoagulation is not achieved until the levels of biologically active factors II (half-life of 50 hours) and X are sufficiently depressed, that is, 4 to 6 days.

Few data exist regarding the risk of spinal hematoma in patients with indwelling spinal or epidural catheters who are subsequently anticoagulated with warfarin. Horlocker et al(19) and Wu and Perkins(20) found no neuraxial hemorrhagic complications in patients who received postoperative epidural analgesia in conjunction with low-dose warfarin after total knee arthroplasty. Because intraspinal hematomas have occurred after removal of the catheter,(6) some have recommended that the same laboratory values apply to placement and removal of an epidural catheter.(21) The current ASRA guidelines recommends an INR value of ≤1.4 as acceptable for the performance of neuraxial blocks.(4) The value was based on studies that showed excellent perioperative hemostasis when the INR value was ≤1.5. The concurrent use of other medications, such as aspirin, NSAIDs, and heparins that affect the clotting mechanism, increases the risk of bleeding complications without affecting the INR.

A controversy exists regarding whether or not the epidural catheter can be removed on postoperative day 1, or 12-14 hours after warfarin is started, when the INR is >1.4. In the absence of other risk factors for increased bleeding, the catheter can probably be removed. The factor VII activity should be determined if risk factors such as low platelets, advanced age, kidney failure, or intake of other anticoagulants are present.(18)

Warfarin is metabolized primarily by the CYP2C9 enzyme of the cytochrome P450 system. Mutations in the gene coding for the cytochrome P450 2C9 hepatic microsomal enzyme affect the elimination clearance of warfarin by impairing the patient’s ability to metabolize S-warfarin. Other genetic factors affecting the warfarin dose-response relationship include polymorphisms of the vitamin K oxide reductase (VKOR) enzyme. Mutations in the gene encoding for isoforms of the protein that can lead to enzymes with varied sensitivities to warfarin is rare, and the American College of Chest Physicians (ACCP) advises against pharmacokinetic-based initial dosing of warfarin at this time.(17)

Intravenous Heparin

Heparin is a complex polysaccharide that exerts its anti- coagulant effect by binding to antithrombin III. The conformational change in antithrombin accelerates its ability to inactivate thrombin, factor Xa, and factor IXa. The anticoagulant effect of subcutaneous heparin takes 1 to 2 hours, but the effect of intravenous heparin is immediate. Heparin has a half-life of 1.5 to 2 hours. The activated partial thromboplastin time (aPTT) is used to monitor the effect of heparin; therapeutic anticoagulation is achieved with a prolongation of the aPTT to >1.5 times the baseline value.

There were no spinal hematomas in >4000 patients who underwent lower extremity vascular surgery under contin- uous spinal or epidural anesthesia.(22) In this study, patients with preexisting coagulation disorders were excluded, heparinization occurred at least 60 minutes after cath- eter placement, the level of anticoagulation was carefully monitored, and the indwelling catheters were removed at a time when heparin activity was low. Ruff and Dougherty (23) noted the occurrence of spinal hematomas in patients who underwent lumbar puncture with subsequent hepa- rinization. The presence of blood during the procedure, concomitant aspirin therapy, and heparinization within 1 hour were identified as risk factors for the development of a spinal hematoma.

When intraoperative anticoagulation is planned, neuraxial technique should be avoided in patients with coexisting coagulopathies. The following considerations are in order:

1. There should be at least a 1-hour delay between needle placement and heparin administration.

2. The catheter should be removed 1 hour before subsequent heparin administration and 2 to 4 hours after the last     heparin dose.(4)

3. The partial thromboplastin time or activated clotting time should be monitored to avoid excessive heparin effect. 

Subcutaneous Heparin

The anticoagulant effect of low molecular weight heparin (LMWH) is similar to unfractionated heparin, that is, activation of antithrombin and acceleration of its interaction with thrombin and factor Xa.(26) LMWH has a greater activity against factor Xa; unfractionated heparin has equivalent activity against thrombin and factor Xa. The plasma half-life of the LMWH ranges from 2 to 4 hours after an intravenous injection and 3 to 6 hours after a subcutaneous injection. Its half-life is two to four times that of standard heparin. The recovery of anti-factor Xa activity after a subcutaneous injection of LMWH approaches 100%. This characteristic makes laboratory monitoring unnecessary, except in patients with renal insufficiency or those with body weight80 kg.

The summary of recommendations for patients receiving LMWH and neuraxial anesthesia are as follows(4):

1. The administration of other anticoagulant medications with LMWHs may increase the risk of spinal hematoma.

2. The presence of blood during needle placement and catheter placement does not necessitate postponement of     surgery. However, the initiation of LMWH therapy should be delayed for 24 hours postoperatively.

3. The first dose of LMWH prophylaxis should be given no earlier than 24 hours postoperatively and only in the     presence of adequate hemostasis.

4. In patients who are on LMWH, needle/catheter placement (or catheter removal) should be performed at least 12     hours after the last prophylactic dose of enoxaparin or 24 hours after higher doses of enoxaparin (1 mg/kg every 12     hours), and 24 hours after dalteparin (120 U/kg every 12 hours or 200 U/kg every 12 hours) or tinzaparin (175 U/kg     daily).

5. The LMWH can be administered 2 hours after the epidural catheter is removed.(4)

6. Monitoring of anti-Xa level is not recommended

Thrombolytic Therapy

Thrombolytic agents actively dissolve fibrin clots that have already formed, secondary to the action of plasmin. Plasminogen activators, such as streptokinase and urokinase, dissolve thrombus and affect circulating plasminogen leading to decreased levels of both plasminogen and fibrin. Clot lysis leads to elevation of fibrin degradation products, which have an anticoagulant effect by inhibiting platelet aggregation. Fibrinogen and plasminogen are maximally depressed at 5 hours after thrombolytic therapy and remain significantly depressed at 27 hours.(4,27)

Although epidural or spinal needle and catheter placement with subsequent heparinization appears relatively safe, the risk of spinal hematoma in patients who receive thrombolytic therapy is less well-defined. Cases of spinal hematoma in patients who received neuraxial injections and thrombolytic agents were reported recently in the medical literature.

Fibrinolytic and thrombolytic agents pose a unique problem when performing neuraxial anesthesia. The time frame for avoidance of these drugs and puncture of noncompressible vessels is 10 days. Except in highly unusual circumstances, patients who received fibrinolytic or thrombolytic drugs should be cautioned against receiving spinal or epidural anesthesia.(4,27) There are no available data to clearly determine the length of time after discontinuation of these drugs for the safe performance of a neuraxial technique. There is no definitive recommendation on the timing of removal of neuraxial catheters in patients who unexpectedly receive fibrinolytic or thrombolytic therapy. Measurement of fibrinogen levels may be helpful in guiding a decision about removal of the catheter.

Herbal Therapy

Herbal preparations do have some effect on platelet aggregation. For example, garlic inhibits platelet aggregation and its effect on hemostasis appears to last 7 days. Ginkgo biloba inhibits platelet-activating factor and its effect lasts 36 hours. These effects last 24 hours with the use of ginseng.(4) The effects of dietary supplements on platelet function and coagulation are not well described, and outcomes are difficult to predict.(28) In spite of these characteristics, herbal preparations appear to present no added significant risk in the development of spinal hematoma in patients having epidural or spinal anesthesia. At this time, there appears to be no specific concerns as to the timing of neuraxial block in relationship to the dosing of herbal therapy, postoperative monitoring, or the timing of neuraxial catheter removal.(4)


Fondaparinux is a synthetic anticoagulant that produces its antithrombotic effect through selective inhibition of factor Xa.(29) The drug exhibits consistency in its anticoagulant effect because it is chemically synthesized and its bioavailability is 100%. Rapidly absorbed, it reaches maximum concentration within 1.7 hours of administration. Its half-life is 17 to 21 hours, allowing once-daily dosing.(30) The actual risk of spinal hematoma with fondaparinux is unknown. The daily dosing makes safe catheter removal harder to predict. The ASRA(4) recommends against the use of fondaparinux in the presence of an indwelling epidural catheter. These recommendations were based on the sustained and irreversible antithrombotic effect of fondaparinux, early postoperative dosing (6 hours after surgery), and the spinal hematoma reported during initial clinical trials. Performance of neuraxial techniques should occur under the conditions used in clinical trials (single needle pass, atraumatic needle placement, and avoidance of indwelling neuraxial catheters).(4)

A 2007 study showed no complications in patients who had neuraxial injections or deep peripheral nerve blocks.(31) In this study, the catheters were removed 36 hours after the last dose of fondaparinux and dosing was delayed for 12 hours after the catheter was removed. In a review article, Rosencher et al(32) recommended that catheter removal should be delayed at least 36 hours (equivalent to two half-lives) and that the subsequent injection should be timed to at least 7 hours after the removal of the catheter.

Thrombin Inhibitors

Recombinant hirudin derivatives, such as desirudin (Revasc), lepirudin (Refludan), and bivalirudin (Angiomax), inhibit both free and clot-bound thrombin.(4) Argatroban, although an L-arginine derivative, is also a thrombin inhibitor. These drugs are used in the treatment of heparin-induced thrombocytopenia and as an adjunct when angioplasty is performed.(33) Their anticoagulant effect is present for 1 to 3 hours after intravenous administration and is monitored by the aPTT. There is no pharmacologic reversal to the effect of these drugs. Desirudin is used as thromboprophylaxis after total hip replacement.(34) There are no published reports of spinal hematoma related to neuraxial anesthesia in patients who have received a thrombin inhibitor, probably because of the hesitancy of clinicians to perform neuraxial injections in patients taking the drugs, which is probably related to their unfamiliarity with the drugs. The most recent ASRA guidelines recommend against the performance of neuraxial techniques in patients who received thrombin inhibitors.

Newer Anticoagulants

Dabigatran Etexilate

Dabigatran is an oral direct thrombin inhibitor. Its bioavailability is only 5%, peak plasma levels occur at 2 hours, and its half-life is 8 hours after a single dose but up to 17 hours after multiple doses. The drug is approved for clinical use in Europe. Studies showed dabigatran (150 or 220 mg daily) to be less effective than enoxaparin (30 mg twice daily) when used for thromboprophylaxis after total joint surgery.(35,36A) 48-hour interval is recommended before a neuraxial injection.


Rivaroxaban is an oral factor Xa inhibitor approved for use in Europe and Canada. It is awaiting approval by the Food and Drug Administration (FDA) in the United States. It has an 80% bioavailability; its peak effect occurs after 1 hour; the duration of effect is 12 hours; and it has a half-life of 9 to 13 hours. Clinical studies comparing rivaroxaban, at doses of 5 to 40 mg, to enoxaparin showed similar or superior efficacy.(37-40) There were no reports of spinal hematoma in these studies. Apparently, a 24-hour interval (2 x half-life) was observed between the rivaroxaban dose and epidural catheter placement or removal; subsequent dosing of the drug was 6 hours after removal of the catheter (personal communication with the company). The drug offers several salutary characteristics including efficacy and simplicity with once-daily oral dosing.


Prasugrel is an oral anticoagulant approved for use by the FDA in July 2009. Its mechanism of action is similar to clopidogrel; that is, it acts as a noncompetitive antagonist of P2Y12, inhibiting the ability of platelet ADP to induce aggregation for the life of the platelet.(41) Prasugrel and clopidogrel are prodrugs; however, prasugrel has a quicker onset of action, a longer duration (the effect of 60 mg is 1-1.5 hours compared with 6 hours with 300 mg clopidogrel); it is 10 times more potent; and less prone to drug-drug interactions and variability in patient response than clopidogrel.(41,42) A 7-10 day interval is recommended before a neuraxial injection. Other novel antiplatelet drugs are in development, including ticagrelor and cangrelor, which are under study for use in patients with acute coronary syndromes.(43)

Anticoagulation and Peripheral Nerve Blocks

Spontaneous hematomas have been reported in patients who took anticoagulants. Abdominal wall hematomas, intracranial hemorrhage, psoas hematoma, and intrahepatic hemorrhage have occurred after LMWH.(44-47) Major hemorrhagic complications occur in 1.9 to 6.5% of patients on enoxaparin.(48) The increased bleeding that occurs after vascular or cardiac procedures and regional nerve blocks in these patients can result in an expanding hematoma with resultant ischemia of the nerve.

There has been no prospective study on peripheral nerve blocks in the presence of anticoagulants. However, there have been several case reports of hematomas when peripheral blocks are performed in patients who are on these drugs. The hematomas occurred in patients with abnormal and normal coagulation status, and in patients who were given LMWH, ticlopidine and clopidogrel, warfarin, heparin, or a combination of the drugs.(49-55) In most cases, however, recovery of neurologic deficits occurred within a year.

The diagnosis of bleeding after peripheral nerve block in patients on anticoagulants include pain (flank, paravertebral, or in the groin with psoas bleeding), tenderness in the area, fall in hemoglobin/hematocrit, fall in blood pressure, and sensory and motor deficits. Although definite diagnosis is made by computed tomography, ultrasound can be a diagnostic aid, and its increasing use will make this modality a useful tool for the diagnosis and subsequent monitoring of peripheral hematomas. Treatment of peripheral hematomas usually includes surgical consult, blood transfusion as necessary, and watchful waiting versus surgical drainage.

The most recent ASRA guidelines recommended that the same guidelines on neuraxial injections apply to deep plexus or peripheral nerve blocks. Some clinicians may find this to be too restrictive and apply the same guidelines only to deep plexus and noncompressible blocks (e.g., lumbar plexus block, deep cervical plexus blocks) or to blocks near vascular areas, such as celiac plexus blocks or superior hypogastric plexus blocks. If peripheral nerve blocks are performed in the presence of anticoagulants, the anesthesiologist must discuss the risks and benefits of the block with the patient and the surgeon, and follow the patient very closely after the block.

Table 1: Summary of Guidelines on Anticoagulants and Neuraxial Blocks*

Guidelines of Various Societies

Guidelines on use of regional anesthesia in the presence of anticoagulants have been published by a number of societ- ies throughout the world to better fit them to the realm of the local practices. By necessity, there are similarities and differences among them. A good example is the new ASRA guidelines (56) and the Belgian and German guidelines.(57,58) The guidelines of the three organizations are similar with regard to antiplatelet medications, unfractionated heparin, and throm- bolytic agents. With regard to LMWH, the ASRA guidelines are more conservative, partly due to the differences in the dosing of the drug. For fondaparinux, the German guidelines allow an indwelling epidural catheter, whereas the ASRA and the Belgian guidelines recommend against it. The Belgian and German guidelines allow neuraxial injections in patients on direct thrombin inhibitors; the ASRA guidelines do not. Finally, some of the newer anticoagulants have been approved for use in Europe and are awaiting approval in the United States so the guidelines for these drugs are forthcoming.


Adherence to the discussed guidelines should lead to a lesser risk of hemorrhagic complications after regional anesthesia, including spinal hematomas. Likewise, implementation of the guidelines leads to improved vigilance and better care of patients on anticoagulants in whom nerve blocks are performed or entertained. Consensus guidelines, however, should be viewed only as recommendations; specific decisions on nerve blocks in patients on anticoagulants should be individualized. Adequate monitoring, follow-up, and timely treatment should be implemented in patients on anticoagulants who are receiving neuraxial or peripheral nerve blocks (see algorithms on the following pages).