Regional Anesthesia in the Anticoagulated Patient:...

Regional Anesthesia in the AnticoagulatedPatient: Defining the Risks (The Second ASRAConsensus Conference on Neuraxial Anesthesiaand Anticoagulation)

Terese T. Horlocker, M.D., Denise J. Wedel, M.D., Honorio Benzon, M.D.,David L. Brown, M.D., F. Kayser Enneking, M.D., John A. Heit, M.D.,Michael F. Mulroy, M.D., Richard W. Rosenquist, M.D., John Rowlingson, M.D.,Michael Tryba, M.D., and Chun-Su Yuan, M.D., Ph.D.

Neuraxial anesthesia and analgesia provide sev-eral advantages over systemic opioids, includ-

ing superior analgesia, reduced blood loss and needfor transfusion, decreased incidence of graft occlu-sion, and improved joint mobility following majorknee surgery.1-4 New challenges in the manage-ment of patients undergoing neuraxial block havearisen over the last 2 decades, as medical standardsfor the prevention of perioperative venous throm-boembolism were established.5,6 Concern for pa-tient safety in the presence of potent antithromboticdrugs has resulted in avoidance of regional anesthe-sia. Indeed, perioperative anesthesia and analgesiaare often determined by the antithrombotic agent.7

Conversely, although the anesthesia community iswell aware of the potential for spinal bleeding,other specialties have only recently become cogni-zant of the risk, as documented by case reports

published in the cardiology and neurology litera-ture.8,9

In response to these patient safety issues, the Amer-ican Society of Regional Anesthesia and Pain Medi-cine (ASRA) convened its Second Consensus Confer-ence on Neuraxial Anesthesia and Anticoagulation.Portions of the material presented here were pub-lished as the proceedings of the 1998 ASRA Consen-sus Conference.10-14 The information has been up-dated to incorporate additional data available sincethe time of its publication. It is important to note thatalthough the consensus statements are based on athorough evaluation of the available information, insome cases data are sparse. Numerous studies havedocumented the safety of neuraxial anesthesia andanalgesia in the anticoagulated patient. Unfortu-nately, with a complication as rare as spinal hema-toma, no clinical study to date has sufficient power todefinitively determine patient management. Conse-quently, the pharmacology of hemostasis-alteringdrugs and case reports of spinal hematoma are alsoessential to regional anesthetic management. Vari-ances from recommendations contained in this docu-ment may be acceptable based on the judgment of theresponsible anesthesiologist. The consensus state-ments are designed to encourage safe and qualitypatient care, but cannot guarantee a specific outcome.They are also subject to timely revision as justified byevolution of information and practice. Finally, thecurrent information focuses on neuraxial blocks andanticoagulants; the risk following plexus and periph-eral techniques remains undefined. Although sev-eral case reports of vascular injury with (or without)resultant nerve dysfunction have been described,15,16

additional experience is needed to allow statementsfor non-neuraxial blocks. The current literature in-volving hemorrhagic complications of plexus and pe-ripheral block is included for completeness.

See Editorial page 163

From the Department of Anesthesiology, Mayo Clinic (T.T.H.,D.J.W., J.A.H.), Rochester, Minnesota; Northwestern University(H.B.), Chicago, Illinois; University of Iowa (D.L.B., R.W.R.), IowaCity, Iowa; University of Florida (F.K.E.), Gainesville, Florida; Vir-ginia Mason Medical Center (M.F.M.), Seattle, Washington; Uni-versity of Virginia Health Science Center (J.R.), Charlottesville,Virginia; Department of Anesthesiology, Intensive Care and PainMedicine (M.T.), Kassel, Germany; and Tang Center for HerbalMedicine Research, Department of Anesthesia and Critical Care,University of Chicago (C-S.Y.), Chicago, Illinois.

Accepted for publication January 29, 2003.Reprint requests: Terese T. Horlocker, M.D., Department of

Anesthesiology, Mayo Clinic, Rochester, MN, 55905. E-mail:[email protected]

© 2003 by the American Society of Regional Anesthesia andPain Medicine.

1098-7339/03/2803-0004$30.00/0doi:10.1053/rapm.2003.50046

172 Regional Anesthesia and Pain Medicine, Vol 28, No 3 (May–June), 2003: pp 172–197

Current Recommendations for thePrevention and Treatment of VenousThromboembolism

Thromboprophylaxis is based upon identificationof risk factors. Guidelines for antithrombotic ther-apy including appropriate pharmacologic agent, de-gree of anticoagulation desired, and duration oftherapy continue to evolve.5,6 Recommendationsfrom the Sixth American College of Chest Physi-cians (ACCP) Consensus Conference in 2001 arebased upon prospective randomized studies that as-sess the efficacy of therapy using contrast venogra-phy or fibrinogen leg scanning to diagnose asymp-tomatic thrombi (Table 1). Clinical outcomes, suchas fatal pulmonary embolism (PE) and symptomaticdeep venous thrombosis (DVT) are not primaryendpoints.17 Despite the successful reduction ofasymptomatic thromboembolic events with routineuse of antithrombotic therapy, an actual reductionof clinically relevant events has been more difficultto demonstrate.18,19 This is in contrast to the docu-mented improvement in perioperative outcomesin selected patient populations that undergoneuraxial anesthesia and analgesia.1-4 Thus, estab-lishment of overall risks and benefits of antithrom-botic therapy in the patient undergoing neuraxialblock is difficult.

Compared with thromboprophylaxis, the pres-ence of acute thromboembolism or unstable angina

necessitates more aggressive antithrombotic (andpotentially thrombolytic) therapy. For example,acute DVT/PE is treated with therapeutic anticoag-ulation with unfractionated or low molecularweight heparin (LMWH); thrombolytic therapymay also be utilized alone or concomitantly (Table1). Likewise, the American College of Cardiologyand American Heart Association Task Force onPractice Guidelines for the Management of Patientswith Acute Myocardial Infarction recommends acombination therapy of (1) aspirin or (2) ticlopi-dine/clopidogrel for patients with aspirin intoler-ance, (3) therapeutic anticoagulation with unfrac-tionated heparin or LMWH, and (4) administrationof a platelet glycoprotein (GP) IIb/IIIa receptor an-tagonist. Intravenous (IV) thrombolytic therapy isrecommended in patients with acute (less than 6hours) symptoms.20 These treatment modalitieshave a dramatic impact on the patient’s ability tomaintain hemostasis. Major bleeding complicationsmay occur spontaneously or at the site of previoustrauma, such as vascular access, surgery, or regionalblock.

Risk of Bleeding Associated WithAntithrombotic and ThrombolyticTherapy

Bleeding is the major complication of anticoagu-lant and thrombolytic therapy. Bleeding is typically

Table 1. Pharmacological Venous Thromboembolism Prophylaxis and Treatment Regimens andTreatment Regimens for Acute Coronary Syndromes

Total Hip or Knee ReplacementThromboprophylaxis

Adjusted-dose unfractionated heparin 3,500 U SC q 8 hours, started 2 hours before surgery; after surgery, thedose is adjusted to maintain the aPTT within the upper normal range

Low molecular weight heparinArdeparin sodium (Normiflow®) 50 U/kg SC q 12 h, started 12-24 hours after surgeryDalteparin sodium (Fragmin®) 5,000 U SC qd, started 12 hours before surgery, or 2,500 U SC given 7

hours after surgery, then 5,000 U SC dailyDanaparoid sodium (Orgaran®) 750 U SC q 12 h, started 2 hours before surgeryEnoxaparin sodium (Lovenox®) 30 mg SC q 12 h, started 12-24 hours after surgery, or 40 mg SC qd,

started 10-12 hours before surgeryTinzaparin (Innohep®) 75 U/kg SC qd, started 10-12 hours before surgery

Warfarin sodium 5 mg orally, started the night before or immediately after surgery andadjusted to prolong the INR � 2.0-3.0

General Surgery ThromboprophylaxisUnfractionated heparin 5,000 U SC q 8-12 hours, started 2 hours before surgeryLow molecular weight heparin

Dalteparin sodium 2,500 U SC qd, started 1-2 hours before surgeryEnoxaparin sodium 40 mg SC qd, started 2 hours before surgery

Acute Coronary Syndrome and VenousThromboembolism Therapy

Enoxaparin sodium 1 mg/kg SC q12 hours (outpatient DVT or non q-wave MI)1 mg/kg SC q12 hours, or 1.5 mg/kg SC qd (inpatient treatment of DVT

or PE)Dalteparin 120 U/kg q 12 hours or 200 U/kg qd (non q-wave MI)Tinzaparin 175 U/kg qd

NOTE. Dosing recommendations from Beerts et al.6Abbreviations: SC, subcutaneous; MI, myocardial infarction; aPTT, activated partial thromboplastin time; INR, international normalized

ratio.

Neuraxial Anesthesia and Anticoagulation • Horlocker et al. 173

classified as major if it is intracranial, intraspinal,intraocular, mediastinal or retroperitoneal, leadsdirectly to death, or results in hospitalization ortransfusion. Risk factors for major bleeding duringanticoagulation include the intensity of the anti-coagulant effect, increased age, female gender, his-tory of gastrointestinal bleeding, concomitant aspi-rin use, and length of therapy.21 During warfarintherapy, an international normalized ratio (INR) of2.0 to 3.0 is associated with a low risk of bleeding:� 3% during a 3-month treatment period. Higherintensity regimens (INR � 4) are associated with asignificantly greater risk of bleeding (7%). The in-cidence of hemorrhagic complications during ther-apeutic anticoagulation with IV or subcutaneous(SC) heparin, as well as LMWH is � 3%.21 Throm-bolytic therapy represents the greatest risk of bleed-ing; between 6% and 30% of patients were treatedwith thrombolytic therapy for DVT.22 Therefore,although thromboembolism remains a source ofsignificant perioperative morbidity and mortality,its prevention and treatment are also associatedwith risk.

Incidence, Risk Factors, and NeurologicOutcome of Spinal Hematoma

Spinal hematoma, defined as symptomatic bleed-ing within the spinal neuraxis, is a rare and poten-tially catastrophic complication of spinal or epiduralanesthesia. The actual incidence of neurologic dys-function resulting from hemorrhagic complicationsassociated with central neural block is unknown. Inan extensive review of the literature, Tryba23 iden-tified 13 cases of spinal hematoma following850,000 epidural anesthetics and 7 cases among650,000 spinal techniques. Based on these observa-tions, the calculated incidence is approximated tobe less than 1 in 150,000 epidurals and less than 1in 220,000 spinal anesthetics.23 Because these esti-mates represent the upper limit of the 95% confi-dence interval, the actual frequency may be muchless. Hemorrhage into the spinal canal most com-monly occurs in the epidural space, most likelybecause of the prominent epidural venous plexus,

although anesthetic variables, such as needle sizeand catheter placement, may also affect the site ofclinically significant bleeding.24,25

In a review of the literature between 1906 and1994, Vandermeulen et al.26 reported 61 cases ofspinal hematoma associated with epidural or spinalanesthesia. In 42 of the 61 patients (68%), thespinal hematomas associated with central neuralblock occurred in patients with evidence of hemo-static abnormality. Twenty-five of the patients hadreceived IV or SC (unfractionated or LMWH), whilean additional 5 patients were presumably adminis-tered heparin as they were undergoing a vascularsurgical procedure. In addition, 12 patients had ev-idence of coagulopathy or thrombocytopenia orwere treated with antiplatelet medications (aspirin,indomethacin, ticlopidine), oral anticoagulants(phenprocoumone), thrombolytics (urokinase), ordextran 70 immediately before or after the spinal orepidural anesthetic. Needle and catheter placementwas reported to be difficult in 15 (25%), or bloodyin 15 (25%) patients. Overall, in 53 of the 61 cases(87%), either a clotting abnormality or needleplacement difficulty was present. A spinal anes-thetic was performed in 15 patients. The remaining46 patients received an epidural anesthetic, includ-ing 32 patients with an indwelling catheter. In 15 ofthese 32 patients, the spinal hematoma occurredimmediately after the removal of the epidural cath-eter. Nine of these catheters were removed duringtherapeutic levels of heparinization. Neurologiccompromise presented as progression of sensory ormotor block (68% of patients) or bowel/bladderdysfunction (8% of patients), not severe radicularback pain. Importantly, although only 38% of pa-tients had partial or good neurologic recovery, spi-nal cord ischemia tended to be reversible in patientswho underwent laminectomy within 8 hours ofonset of neurologic dysfunction26 (Table 2).

The need for prompt diagnosis and interventionin the event of a spinal hematoma was also dem-onstrated in a recent review of the American Soci-ety of Anesthesiologists (ASA) Closed Claims data-base, which noted that spinal cord injuries were the

Table 2. Neurologic Outcome in Patients With Spinal Hematoma Following Neuraxial Block

Interval Between Onsetof Paraplegia and Surgery

GoodN � 15

PartialN � 11

PoorN � 29

Less than 8 hours (N � 13) 6 4 3Between 8 and 24 hours (N � 7) 1 2 4Greater than 24 hours (N � 12) 2 0 10No surgical intervention (N � 13) 4 1 8Unknown (N � 10) 2 4 4

NOTE. Neurologic outcome was reported for 55 of 61 cases of spinal hematoma following neuraxial block.Adapted and reprinted with permission.26

174 Regional Anesthesia and Pain Medicine Vol. 28 No. 3 May–June 2003

leading cause of claims in the 1990s.27 Spinal he-matomas accounted for nearly half of the spinalcord injuries. Risk factors for spinal hematoma in-cluded epidural anesthesia in the presence of IVheparin during a vascular surgical or diagnostic pro-cedure. Importantly, the presence of postoperativenumbness or weakness was typically attributed tolocal anesthetic effect rather than spinal cord ische-mia, which delayed the diagnosis. Patient care wasrarely judged to have met standards (1 of 13 cases)and the median payment was very high.

Fibrinolytic and Thrombolytic Therapy

Pharmacology of Fibrinolytics/Thrombolytics

The fibrinolytic system dissolves intravascularclots as a result of the action of plasmin. Plasmin isproduced by the cleavage of a single peptide bond ofthe inactive precursor, plasminogen. The resultingcompound is a nonspecific protease capable of dis-solving fibrin clots and other plasma proteins, in-cluding several coagulation factors. Exogenousplasminogen activators, such as streptokinase andurokinase, not only dissolve thrombus, but alsoaffect circulating plasminogen as well. Endogenoust-PA formulations (alteplase and tenecteplase) aremore fibrin-selective and have less effect on circu-lating plasminogen. Clot lysis leads to elevation offibrin degradation products, which themselves havean anticoagulant effect by inhibiting platelet aggre-gation. In addition to the fibrinolytic agent, thesepatients frequently receive IV heparin to maintainan activated partial thromboplastin time (aPTT) of1.5 to 2 times normal, and often an antiplateletagent, such as aspirin or clopidogrel. While theplasma half-life of thrombolytic drugs is only hours,it may take days for the thrombolytic effect to re-solve; fibrinogen and plasminogen are maximallydepressed at 5 hours after thrombolytic therapy andremain significantly depressed at 27 hours. The de-crease in coagulation factor levels is greater withstreptokinase compared with t-PA therapy. How-ever, the frequency of hemorrhagic events is simi-lar.28 Importantly, original contraindications tothrombolytic therapy included surgery or punctureof noncompressible vessels within 10 days.28

Case Reports of Spontaneous and RegionalAnesthesia-Related Spinal Hematomas Relatedto Thrombolytic Therapy

There are no published studies addressing spinal,epidural, or regional anesthesia in the patient re-ceiving fibrinolytic/thrombolytic therapy. However,there is limited information about these settingsavailable in the form of case reports of spinal he-

matoma. The majority of published reports involvespontaneous spinal or epidural hematomas afterthrombolytic therapy.29-35 To date, there are 5 casesof spinal hematoma involving the concomitant useof neuraxial anesthesia and fibrinolytic/thrombo-lytic therapy. Four cases appeared in the litera-ture8,36-38; 1 additional case was reported throughthe MedWatch1 system. An epidural technique hadbeen performed in 3 patients, a continuous spinalanesthetic in 1 patient, and an epidural steroid in-jection in the remaining patient. In 3 of the cases,the patients presented with lower extremity ische-mia, and a neuraxial anesthetic was performed toallow surgical revascularization. However, 2 of therecent spinal hematomas (including the MedWatchcase) occurred in patients who underwent aneuraxial technique (epidural anesthesia for litho-tripsy, epidural steroid injection2) and subsequentlycomplained of myocardial ischemia and weretreated with a thrombolytic.8 The potential for sig-nificant spinal bleeding was not appreciated despiterecent neuraxial needle placement in these 2 pa-tients.

Anesthetic Management of the Patient ReceivingThrombolytic Therapy

Patients receiving fibrinolytic/thrombolytic med-ications are at risk of serious hemorrhagic events,particularly those who have undergone an invasiveprocedure. Consensus statements are based on theprofound effect on hemostasis, the use of concom-itant heparin and/or antiplatelet agents (which fur-ther increase the risk of bleeding), and the potentialfor spontaneous neuraxial bleeding with thesemedications.

Advances in fibrinolytic/thrombolytic therapyhave been associated with an increased use of thesedrugs, which will require further increases in vigi-lance. Ideally, the patient should be queried prior tothe thrombolytic therapy for a recent history oflumbar puncture, spinal or epidural anesthesia, orepidural steroid injection to allow appropriate mon-itoring. Guidelines detailing original contraindica-

1 The MedWatch program was initiated in 1993. Reporting ofserious adverse events by health care professionals and hospitalsis voluntary. Confidentiality is maintained. However, manufac-turers and distributors of FDA-approved pharmaceuticals havemandatory reporting requirements. The FDA estimates that lessthan 1% of serious adverse drug reactions are reported (Gold-man, 1996).

2 An 84-year-old male received an uncomplicated epiduralsteroid injection in the morning. He developed chest pain laterthat day, was admitted to the hospital, diagnosed with an acutemyocardial infarction (MI), and treated with t-PA and heparin.He subsequently developed back pain and paraplegia. MRI dem-onstrated an epidural hematoma extending from T10 to thesacrum. Treatment and outcome were not reported.


tions for thrombolytic drugs suggest avoidance ofthese drugs for 10 days following puncture of non-compressible vessels.

Preoperative evaluation should determinewhether fibrinolytic or thrombolytic drugs havebeen used preoperatively or have the likelihood ofbeing used intraoperatively or postoperatively. Pa-tients receiving fibrinolytic and thrombolytic drugsshould be cautioned against receiving spinal or epi-dural anesthetics except in highly unusual circum-stances. Data are not available to clearly outlinethe length of time neuraxial puncture should beavoided after discontinuation of these drugs.

In those patients who have received neuraxialblocks at or near the time of fibrinolytic and throm-bolytic therapy, neurologic monitoring should becontinued for an appropriate interval. It may bethat the interval of monitoring should not be morethan 2 hours between neurologic checks. Further-more, if neuraxial blocks have been combined withfibrinolytic and thrombolytic therapy and ongoingepidural catheter infusion, the infusion should belimited to drugs minimizing sensory and motorblock to facilitate assessment of neurologic func-tion.

There is no definitive recommendation for re-moval of neuraxial catheters in patients who unex-pectedly receive fibrinolytic and thrombolytic ther-apy during a neuraxial catheter infusion. Themeasurement of fibrinogen level (one of the lastclotting factors to recover) may be helpful in mak-ing a decision about catheter removal or mainte-nance.

Unfractionated IV and SC Heparin

Pharmacology of Unfractionated Heparin

The major anticoagulant effect of heparin is dueto a unique pentasaccharide that binds to anti-thrombin (AT) with high affinity and is present inapproximately one third of heparin molecules.Binding of this heparin pentasaccharide to AT ac-celerates its ability to inactivate thrombin (factorIIa), factor Xa, and factor IXa. Anticoagulant activ-ities of unfractionated heparin depend on both thenumber of heparin molecules with the pentasac-charide chain and the size of the molecules contain-ing the pentasaccharide sequence. Larger molecularweight heparins will catalyze inhibition of both fac-tor IIa and Xa. Smaller molecular weight heparinswill catalyze inhibition of only factor Xa.39 IV injec-tion results in immediate anticoagulant activity,whereas SC injection results in a 1- to 2-hour delay.The anticoagulant effect of heparin is both dose-and molecular size-dependent and is not linear, butincreases disproportionately with increasing doses.

For example, the biologic half-life of heparin in-creases from 30 minutes after 25 U/kg IV to 60minutes with 100 U/kg IV.39

The anticoagulant effect of heparin is typicallymonitored with the aPTT. Adequate therapeutic ef-fect (in patients with venous thromboembolism orunstable angina) is achieved with a prolongation ofthe aPTT to greater than 1.5 times the baselinevalue, heparin level of 0.2 to 0.4 U/mL or anti-Xalevel of 0.3 to 0.7 U/mL.40 Administration of smalldose (5,000 U) SC heparin for prophylaxis of deepvenous thrombosis generally does not prolong theaPTT, and is typically not monitored. It can result inunpredictable (10-fold variability) and therapeuticblood concentrations of heparin in some patientswithin 2 hours after administration.41

Risk Factors for Spinal Hematoma in theHeparinized Patient Undergoing Neuraxial Block

The combination of spinal or epidural needle in-sertion in the presence of anticoagulation with hep-arin may be associated with increased risk. Much ofour information about this association comes froma report of 342 patients who deliberately receivedsystemic therapeutic heparin after lumbar punc-ture.42 Until the routine use of computed tomogra-phy (CT) in the 1980s, diagnostic subarachnoidpuncture was routinely used to select patients forheparin therapy for acute cerebral ischemia. Ruffand Dougherty42 reported that 7 of 342 patientstreated in this manner developed spinal hemato-mas. Three factors associated with increased riskwere identified: � 60-minute time interval betweenthe administration of heparin and lumbar puncture,traumatic needle placement, and concomitant useof other anticoagulants (aspirin). These risk factorshave been verified in subsequent large reviewsof case reports of hematomas associated withneuraxial procedures in the presence of unfraction-ated heparin (Table 3).25 In addition, the resultshave been utilized to define safe practice protocolsfor patients undergoing neuraxial block during sys-temic heparinization, particularly during vascularsurgery.26,43

Intraoperative Systemic Heparinization

Intraoperative heparinization typically involvesIV injection of 5 to 10,000 U heparin during theoperative period, particularly in the setting of vas-cular surgery to prevent coagulation during cross-clamping of arterial vessels.39 Neuraxial anesthetictechniques are often attractive for these patients, asthese techniques may provide reduced morbidityand improved postoperative analgesia.1 However,the use of neuraxial procedures in the presence of


unfractionated heparin may be associated with anincreased risk of epidural hematoma.26,27

Most published case series used similar guidelinesfor patient management, including exclusion ofhigh-risk patients (preexisting coagulopathy) andperformance of neuraxial procedure at least 1 hourprior to administration of heparin.13 The questionof how to manage the situation of a bloody ortraumatic neuraxial procedure has been raised. Pre-vious case reports suggest that presence of a bloodytap or a traumatic regional block is an associatedfactor in approximately 50% of spinal hemato-mas.26 Although some investigators have recom-mended cancellation of the surgical proceduresshould these events occur,43 there are no clinicaldata to support this recommendation.44,45 Directcommunication with the surgeon and a specificrisk-benefit decision about proceeding in each caseis warranted.

Heparinization may be continued into the post-operative period. Prolonged IV heparin administra-tion is usually performed with a constant IV infu-sion of heparin, usually with a goal of aPTTprolongation to 1.5 to 2 times the baseline level.The risk of any (spontaneous, surgical, or anesthe-sia-related) bleeding due to heparin in such an an-ticoagulated patient may be increased.21 Most im-portantly, the initiation of systemic therapeuticheparin therapy for medical or surgical indicationsin the presence of a neuraxial catheter potentiallyincreases the risk of hematoma formation duringcatheter removal. In the series by Vandermeulen etal.,26 half of the spinal hematomas associated withsystemic heparinization occurred at the time ofcatheter removal. In all patients who have under-gone systemic heparinization, the heparin shouldbe discontinued for 2 to 4 hours prior to neuraxialcatheter removal, coagulation status assessed beforemanipulation of the catheter, and careful assess-ment of the presence of sensory and motor functionin the lower extremities for at least 12 hours fol-lowing the catheter removal.

Overall, large published series and extensive clin-ical experience suggests the use of regional tech-niques during systemic heparinization does not ap-pear to represent a significant risk.13 However, therecent reports of paralysis relating to spinal hema-toma in the ASA Closed Claims database suggeststhat these events may not be as rare as suspectedand that vigilance is necessary to diagnose and in-tervene as early as possible, should spinal hema-toma be suspected.27

Complete Anticoagulation duringCardiopulmonary Bypass

Since the publication of the ASRA ConsensusConference guidelines in 1998,13 there have beencontinued discussions regarding the relative risk(and benefit) of neuraxial anesthesia and analgesiain the patient undergoing heparinization for cardio-pulmonary bypass. To date, there are no cases ofspinal hematoma associated with this techniquepublished or within the Closed Claims Project.27 Areview has recommended certain precautions to betaken46: (1) neuraxial blocks should be avoided in apatient with known coagulopathy from any cause;(2) surgery should be delayed 24 hours in the eventof a traumatic tap; (3) time from instrumentation tosystemic heparinization should exceed 60 minutes;(4) heparin effect and reversal should be tightlycontrolled (smallest amount of heparin for theshortest duration compatible with therapeutic ob-jectives); and (5) epidural catheters should be re-moved when normal coagulation is restored, andpatients should be closely monitored postopera-tively for signs and symptoms of hematoma forma-tion.

These recommendations have been used by mostof the published case series, as well as the practice ofinserting epidural catheters 24 hours in advance ofsurgery. Validity of these and future recommenda-tions will need to be determined as experience withneuraxial procedures in these patients increases. In

Table 3. Risk Factors and Estimated Incidence for Spinal Hematoma and Central Neuraxial Anesthesia

Relative Risk ofSpinal Hematoma

Estimated Incidence forEpidural Anesthesia

Estimated Incidence forSpinal Anesthesia

No heparinAtraumatic 1.00 1:220,000 1:320,000Traumatic 11.2 1:20,000 1:29,000With aspirin 2.54 1:150,000 1:220,000

Heparin anticoagulation following neuraxial procedureAtraumatic 3.16 1:70,000 1:100,000Traumatic 112 1:2,000 1:2,900Heparin � 1 hr after puncture 2.18 1:100,000 1:150,000Heparin � 1 hr after puncture 25.2 1:8,700 1:13,000With aspirin 26 1:8,500 1:12,000

Reprinted with permission.25


a recent survey of the membership of the Society ofCardiovascular Anesthesiologists, Goldstein et al.47

surveyed 3,974 cardiac anesthesiologists and found7% of their responders used spinal or epidural tech-niques for cardiac surgery. Interestingly, the major-ity of anesthesiologists would proceed if frank bloodwere noted in the spinal or epidural needle.

At the time of the original consensus conference,there were not felt to be enough available data toidentify the specific risk associated with this tech-nique.13,46 Since that time, further reports of smallseries have appeared, again with no reported com-plications. Two of these series are retrospective re-views of pediatric cardiac surgery including a totalof 250 patients that report no spinal hemato-mas.48,49 In these pediatric patients the blocks wereperformed after induction of general anesthesiaprior to surgery 1 hour before full systemic hepa-rinization. In contrast, the adult experience withcoronary bypass surgery has continued to followthe practice of placement of the epidural catheterson the evening prior to surgery. Sanchez and Ny-gard50 report a large prospective series of 558 pa-tients without complications. Despite the absence ofserious sequelae, the debate continues as to therisk-benefit advantages of this technique.51,52 Re-cently, the efficacy has been examined in the newer“off-pump” approach to cardiac surgery.53,54 In aseries of 50 patients, Priestly et al.55 reported im-proved postoperative analgesia and earlier extuba-tion. However, there was no difference in time tohospital dismissal.55 Although there were no spinalhematomas, the authors observe that “the use ofthoracic epidural analgesia during coronary arterybypass grafting is controversial because the anti-coagulation required during surgery raises the con-cern of increasing the rare but serious risk of per-manent spinal cord damage from an epiduralhematoma. Such a risk must be balanced by impor-tant clinical advantages if the technique is to bejustified.” Despite improved analgesia, they notethat “convincing respiratory, cardiac, or other organoutcome data are lacking.”

Another approach to this dilemma is presentedby Ho et al. who calculated the risk of hematoma. Ina complex mathematical analysis of the probabilityof predicting a rare event that has not occurred yet,they estimate the probability of a spinal hematoma(based on the totals of 4,583 epidural and 10,840spinal anesthetics reported without complications)to be in the neighborhood of 1:1,528 for epiduraland 1:3,610 for spinal technique.56 The authorshypothesized that this may be an acceptable riskcompared with the mortality of postoperative myo-cardial infarction in this high-risk population, butconcede that the actual risk of hematoma may be

higher with broader usage, and that further studiesare certainly needed. It is of interest that most au-thors recognize the risks associated with the imme-diate perioperative period and are placing indwell-ing catheters the night before surgery. This maybecome an increasingly difficult aspect of a practicewhen the majority of cardiac patients are now alsobeing treated as “morning admission” patients inmost hospitals. The future of neuraxial anesthesiaand analgesia for coronary bypass surgery remainssomewhat unclear.57

Low-Dose SC Heparin

Low-dose heparin is commonly used for prophy-laxis against development of venous thromboem-bolism in general and urologic surgery.58 SC admin-istration of 5,000 U heparin every 12 hours hasbeen used extensively and effectively for prophy-laxis against deep venous thrombosis. There is oftenno detectable change in the clotting parameters, asmeasured by the aPTT. There are a minority ofpatients, perhaps up to 15%, who may developmeasurable changes in coagulation, although theaPTT rarely exceeds 1.5 times the normal level.41

There is a smaller subset (2% to 4%) of patientswho may become therapeutically anticoagulatedduring SC heparin therapy. With therapy greaterthan 5 days, there is a subset of patients who willdevelop a decrease in the platelet count.39

The widespread use of SC heparin and paucity ofcomplications suggest that there is little risk of spi-nal hematoma associated with this therapy. Thereare 9 published series totaling over 9,000 patientswho have received this therapy without complica-tions,13 as well as extensive experience in both Eu-rope and United States without a significant fre-quency of complications. Three surveys of opinionsamong anesthesiologists in Denmark,59 Great Brit-ain, and Scotland,60 and New Zealand61 report thatthe majority of anesthesiologists appear to feel thatthe presence of SC heparin prophylaxis is not astrong contraindication to the performance ofneuraxial anesthesia. There are only 4 case reportsof neuraxial hematomas, 3 epidural26 and 1 sub-arachnoid,62 during neuraxial block with the use ofSC heparin.

Performance of neuraxial block before the injec-tion of SC heparin may be preferable, but theredoes not appear to be an increased risk withneuraxial block in the presence of SC heparin. Pre-vious investigators have recommended delayingperformance of neuraxial blocks for 2 hours afteradministration of SC heparin.44 However, this mayactually coincide with peak effect, and clinical ex-perience questions the need for this delay.


Since the time of our first consensus conference,one additional spinal hematoma has been reportedfollowing epidural catheter placement in a patientreceiving SC heparin.63 The patient had severalother risk factors involved, and guidelines for inser-tion and removal of catheters presented in the pre-vious consensus publication13 were not followed.Nevertheless, this confirms that there is a small butvery limited risk associated with the use of epiduraland spinal anesthesia in the presence of SC heparintreatment.

Other areas of heparin therapy have not ap-peared to increase the incidence of spinal hema-toma, at least based upon the case reports in theliterature and reports to the MedWatch system.Based on this apparent frequency, there does notappear to be any reason to significantly modify theinitial consensus statements.

Anesthetic Management of the Patient ReceivingUnfractionated Heparin

Anesthetic management of the heparinized pa-tient was established over 2 decades ago. Initialrecommendations have been supported by in-depthreviews of case series, case reports of spinal hema-toma, and the ASA Closed Claims Project.

During SC (mini-dose) prophylaxis, there is nocontraindication to the use of neuraxial techniques.The risk of neuraxial bleeding may be reduced bydelay of the heparin injection until after the blockand may be increased in debilitated patients afterprolonged therapy. Because heparin-inducedthrombocytopenia may occur during heparin ad-ministration, patients receiving heparin for greaterthan 4 days should have a platelet count assessedprior to neuraxial block and catheter removal.

Combining neuraxial techniques with intraoper-ative anticoagulation with heparin during vascularsurgery seems acceptable with the following cau-tions: (1) avoid the technique in patients with othercoagulopathies; (2) heparin administration shouldbe delayed for 1 hour after needle placement; (3)indwelling neuraxial catheters should be removed 2to 4 hours after the last heparin dose and the pa-tient’s coagulation status is evaluated and rehepa-rinization should occur 1 hour after catheter re-moval; (4) monitor the patient postoperatively toprovide early detection of motor block and consideruse of minimal concentration of local anesthetics toenhance the early detection of a spinal hematoma;(5) although the occurrence of a bloody or difficultneuraxial needle placement may increase risk,there are no data to support mandatory cancellationof a case. Direct communication with the surgeon

and a specific risk-benefit decision about proceed-ing in each case is warranted.

Currently, insufficient data and experience areavailable to determine if the risk of neuraxial he-matoma is increased when combining neuraxialtechniques with the full anticoagulation of cardiacsurgery. Postoperative monitoring of neurologicfunction and selection of neuraxial solutions thatminimize sensory and motor block is recommendedto facilitate detection of new/progressive neurodefi-cits.

The concurrent use of medications that affectother components of the clotting mechanisms mayincrease the risk of bleeding complications for pa-tients receiving standard heparin. These medica-tions include antiplatelet medications, LMWH, andoral anticoagulants.

LMWH

Pharmacology, Monitoring, and Reversal of theAnticoagulant Effect of LMWH

The biochemical and pharmacologic properties ofLMWH differ from those of unfractionated hepa-rin.64-67 Most relevant are the lack of monitoring ofthe anticoagulant response (anti-Xa level), pro-longed half-life, and irreversibility with protamine.Prolonged LMWH therapy may be associated withan accumulation of anti-Xa activity and fibrinoly-sis.68 In addition, the plasma half-life of LMWHincreases in patients with renal failure.64 LMWHsvary both biochemically and pharmacologically, in-cluding molecular weight, anti-IIa and anti-Xa ac-tivities, and plasma half-life. However, becausethere are no adequate trials comparing the efficacyand safety of one LMWH to another, it is impossibleto recommend one specific LMWH over another.Experience in Europe suggests that the rate of spi-nal hematoma is similar among LMWH prepara-tions.69

Spinal and Epidural Anesthesia in the PatientReceiving LMWH

The relative rarity of spinal hematoma, as well asthe publication of several large studies reportedover the last 2 decades, increased the clinician’sconfidence in management of the anticoagulatedpatient undergoing neuraxial block. However, anew challenge occurred with the release of LMWHfor general use in the United States in May 1993.The pharmacologic differences between LMWH andstandard heparin were underestimated; over 40 spi-nal hematomas were reported through the Med-Watch system over a 5-year period. This experiencecontrasted dramatically with that of European cli-


nicians who had reported only 13 spinal hemato-mas in patients receiving LMWH despite a decade ofextensive clinical use.14 The marked increase in thefrequency of spinal hematoma in patients anticoag-ulated with LMWH prompted a reevaluation of therelative risks and benefits of neuraxial block.

LMWH Thromboprophylaxis in Europe

The administration of LMWH in patients under-going spinal or epidural anesthesia was examinedby Bergqvist et al.70,71 in 2 reviews published in1992 and 1993. These studies represent the Euro-pean experience with LMWH thromboprophylaxis,since no LMWH preparation had been approved forgeneral use in the United States at that time.Bergqvist et al.72 identified 19 articles involving9,013 patients who safely received the combinationof LMWH and spinal or epidural anesthesia. Impor-tantly, only 1 case of spinal hematoma had beenreported.72 To further document the safety ofLMWH in combination with neuraxial block, theauthors noted that although only 9,013 patientswere identified in their review, pharmaceuticalcompanies had estimated that several million pa-tients had received LMWH while undergoingneuraxial techniques. Based on these data,Bergqvist et al.71 concluded that neurologic compli-cations after spinal or epidural anesthesia in pa-tients receiving LMWH thromboprophylaxis are ex-tremely rare, and the combination appeared safe.However, a careful analysis of European literaturebetween the years 1993 and 1995 reveals that con-sistent practice guidelines among the European so-cieties had been established.73,74 The recommenda-tions established time intervals between LMWHdosing and needle placement/catheter removal, aswell as guidelines for neurologic monitoring.75

These guidelines were apparently effective in re-ducing the frequency of spinal hematoma in pa-tients receiving the combination of regional anes-thesia and LMWH; only 13 spinal hematomas werereported among the European community between1989 and 1998.14 In addition, epidural analgesia isnot considered a contraindication to concomitantLMWH thromboprophylaxis in Europe.76,77 Itshould be noted that European dosing of LMWH isonce daily, with the first dose administered 10 to 12hours preoperatively.

LMWH Thromboprophylaxis in North America

Enoxaparin (Aventis Pharmaceuticals Products,Inc., Bridgewater, NJ), the first LMWH to be ap-proved by the Food and Drug Administration (FDA)in the United States, was distributed for general usein May 1993. At that time, labeled indications in-

cluded thromboprophylaxis after major joint re-placement. Approved dose scheduling was 30 mgevery 12 hours, with the first dose administered assoon as possible after surgery. Within 1 year, 2 casesof spinal hematoma had been voluntarily reportedthrough the MedWatch system. The WARNINGS sec-tion of the drug label was revised in March 1995 tocaution practitioners of the risk of spinal hematomain patients with indwelling catheters or concomi-tant treatment with antiplatelet medications, andthe prescribing information was changed to recom-mend that the first dose be given 12 to 24 hoursafter surgery (rather than immediately postopera-tively).

Despite repeated efforts at relabeling and educa-tion, cases of spinal hematoma continued to occur.An FDA Health Advisory was issued in December1997. In addition, the manufacturers of all LMWHand heparinoids were requested to include a black“boxed warning,” a designation used by the FDA todenote adverse events, particularly those that maylead to death or serious injury (Code of FederalRegulations, last updated July 15, 2001), in thelabeling of their respective products.

Risk Factors for Spinal Hematoma

A review and update of the LMWH literature wasperformed in 1997 to determine the incidence andrisk factors associated with spinal hematoma, aswell as practice differences between Europe andNorth America.78 Thirty-nine studies involving15,151 neuraxial anesthetics performed in combi-nation with perioperative LMWH thromboprophy-laxis were identified. Although it was documentedthat a single dose spinal was performed in nearlyhalf of the patients included, it was impossible todetermine the actual number of continuous epi-dural anesthetics (and the duration of postoperativeanalgesia) because the anesthetic technique wasoften recorded as “spinal or epidural” or “regionalanesthesia.” It is notable that none of the 30 cases ofspinal hematoma had occurred within this series.

At the time of the Consensus Conference onNeuraxial Anesthesia and Anticoagulation in April1998, there were 45 cases of spinal hematoma as-sociated with LMWH, 40 of which involved aneuraxial anesthetic. Severe radicular back painwas not the presenting symptom; most patientscomplained of new onset numbness, weakness, orbowel and bladder dysfunction. Approximately halfof patients undergoing a continuous technique re-ported neurologic deficits 12 hours or more follow-ing catheter removal. Median time interval be-tween initiation of LMWH therapy and neurologicdysfunction was 3 days, while median time to onset


of symptoms and laminectomy was over 24 hours.Less than one third of the patients reported fair orgood neurologic recovery.14

The risk of spinal hematoma (based on LMWHsales, prevalence of neuraxial techniques, and re-ported cases) was estimated to be approximately 1in 3,000 continuous epidural anesthetics comparedwith 1 in 40,000 spinal anesthetics.79 However, thisis most likely an underestimation. In addition to thespinal hematomas that had been reported at thetime of the First Consensus Conference, there wereapproximately 20 more that had occurred, but werenot yet reported to the MedWatch system. In total,nearly 60 spinal hematomas were tallied by theFDA between 1993 and 1998.

Based on an examination of the published cases,MedWatch reports, and clinical experience in Eu-rope and North America, specific risk factors havebeen proposed. It is not possible to stratify the in-dividual risk factors or determine interactions be-tween risk factors (Table 4). A recent clinical inves-tigation has reported a significant anticoagulanteffect present at the time of epidural catheter re-moval in patients receiving twice daily LMWH com-pared with once daily LMWH administration.80

Reports of Spinal Hematoma Since 1998

Since 1998, there have been 6 cases of spontane-ous spinal hematomas associated with LMWH, in-cluding 1 patient who also received a thrombolytic.These cases represent the newer applications ofLMWH, which involve treatment of DVT/PE andacute coronary syndrome (Table 1).

There have been only 13 cases of spinal hema-toma following neuraxial block since 1998 reportedthrough the MedWatch system or published as casereports.81,82 Five cases were from outside the UnitedStates, including the 2 published case reports. Inaddition to LMWH, 5 patients received ketorolac, 1patient received ibuprofen,82 and 1 patient receivedIV unfractionated heparin during a vascular proce-

dure. The regional technique was a spinal anes-thetic in 3 cases. The remaining 10 patients under-went epidural anesthesia in combination withLWMH therapy. Thus, the characteristics of the re-ported cases support the previous recommenda-tions of epidural catheter removal prior to theinitiation of LMWH thromboprophylaxis and avoid-ance of concomitant antiplatelet/anticoagulantmedications. Although the number of cases volun-tarily reported has markedly declined, this may be aresult of decreased reporting, improved manage-ment, or simple avoidance of all neuraxial tech-niques in patients receiving LMWH. Continuedmonitoring is necessary.

New Applications of LMWH

The indications and labeled uses for LMWH con-tinue to evolve. Indications for thromboprophylaxisas well as treatment of DVT/PE or MI have beenintroduced since the first Consensus Conference.These new applications and corresponding regionalanesthetic management warrant discussion. Addi-tional research has been performed regarding theefficacy of a single daily dose of LMWH for throm-boprophylaxis following major joint surgery. Previ-ous labeling had included a once daily dose of enox-aparin 40 mg; however, the first dose wasadministered 12 hours preoperatively. Since thisnecessitated hospital admission the night beforesurgery, it was seldom utilized. Rather, the twice-daily dosing regimen of enoxaparin 30 mg every 12hours, with the first dose administered postopera-tively, was most common. Recently, the efficacy ofsingle daily administration, with postoperative ini-tiation of therapy, has been approved for Dalteparin(Pharmacia and Upjohn Company, Kalamazoo,MI).83 This application involves administration ofhalf the usual dose (2,500 U) approximately 6 to 8hours postoperatively. The next dose (5,000 U) isgiven 24 hours later. This regimen approximatesthe European dosing schedule, but without the pre-operative dose. Thus, it should be possible to man-age patients according to European guidelines.Indwelling epidural catheters may be safely main-tained with this administration. However, timing ofboth the first and subsequent doses of dalteparinshould be confirmed to assure an adequate timeinterval between LMWH doses as well as needleplacement/catheter removal.

Several off-label applications of LMWH are ofspecial interest to the anesthesiologist. LMWH hasbeen demonstrated to be efficacious as a “bridgetherapy” for patients chronically anticoagulatedwith warfarin, including parturients, patients withprosthetic cardiac valves, a history of atrial fibrilla-

Table 4. Patient, Anesthetic, and LMWH DosingVariables Associated With Spinal Hematoma

Patient factorsFemale genderIncreased age

Anesthetic factorsTraumatic needle/catheter placementEpidural (compared to spinal) techniqueIndwelling epidural catheter during LMWH administration

LMWH dosing factorsImmediate preoperative (or intraoperative) LMWH

administrationEarly postoperative LMWH administrationConcomitant antiplatelet or anticoagulant medicationsTwice daily LMWH administration


tion, or a preexisting hypercoagulable condition.84

In anticipation of surgery, warfarin is discontinuedand the prothrombin time allowed to normalize.During this time, the patient would be at risk forthromboembolic events, and historically would behospitalized and heparinized systemically. Outpa-tient LMWH is a suitable alternative. The doses ofLMWH are those associated with DVT treatment,not prophylaxis, and are much higher (Table 1).Needle placement should occur a minimum of 24hours following this level of LMWH anticoagula-tion. It is also important to determine when the firstpostoperative dose is anticipated, because these pa-tients are often aggressively anticoagulated postop-eratively. In these cases, a spinal or a general anes-thetic may be the safest alternatives.

Efficacy of Management Guidelines in Reducingthe Risk of Spinal Hematoma

Perioperative management of patients receivingLMWH requires coordination and communication.Time intervals between neuraxial needle placementand administration of LMWH must be maintained.However, hospital staffs often administer LMWH ata set time (usually 7 to 8 AM and 7 to 8 PM), unlessotherwise specified. It is also important to note thateven when protocols for dosing of LMWH and cath-eter management exist, they may not be closelyfollowed. McEvoy et al.85 reported a 52% noncom-pliance rate in the administration of LMWH in as-sociation with epidural analgesia. Clinicians areurged to develop protocols that “fit” within thenormal practice standards at their institutionsrather than deviate from the routine.

Anesthetic Management of the Patient ReceivingLMWH

Anesthesiologists in North America can draw onthe extensive European experience to develop prac-tice guidelines for the management of patients un-dergoing spinal and epidural blocks while receivingperioperative LMWH. All consensus statementscontained herein respect the labeled dosing regi-mens of LMWH as established by the FDA. Al-though it is impossible to devise recommendationsthat will completely eliminate the risk of spinalhematoma, previous consensus recommendationshave appeared to improve outcome. Concern re-mains for higher dose applications, where sustainedtherapeutic levels of anticoagulation are present.

Monitoring of the anti-Xa level is not recom-mended. The anti-Xa level is not predictive of therisk of bleeding and is, therefore, not helpful in themanagement of patients undergoing neuraxialblocks.

Antiplatelet or oral anticoagulant medicationsadministered in combination with LMWH may in-crease the risk of spinal hematoma. Concomitantadministration of medications affecting hemostasis,such as antiplatelet drugs, standard heparin, or dex-tran, represents an additional risk of hemorrhagiccomplications perioperatively, including spinal he-matoma. Education of the entire patient care teamis necessary to avoid potentiation of the anticoagu-lant effects.

The presence of blood during needle and catheterplacement does not necessitate postponement ofsurgery. However, initiation of LMWH therapy inthis setting should be delayed for 24 hours postop-eratively. Traumatic needle or catheter placementmay signify an increased risk of spinal hematoma,and it is recommended that this consideration bediscussed with the surgeon.

Preoperative LMWH

Patients on preoperative LMWH thrombopro-phylaxis can be assumed to have altered coagula-tion. In these patients, needle placement shouldoccur at least 10 to 12 hours after the LMWH dose.

Patients receiving higher (treatment) doses ofLMWH, such as enoxaparin 1 mg/kg every 12hours, enoxaparin 1.5 mg/kg daily, dalteparin 120U/kg every 12 hours, dalteparin 200 U/kg daily, ortinzaparin 175 U/kg daily will require delays of atleast 24 hours to assure normal hemostasis at thetime of needle insertion.

Neuraxial techniques should be avoided inpatients administered a dose of LMWH 2 hourspreoperatively (general surgery patients), becauseneedle placement would occur during peak antico-agulant activity.

Postoperative LMWH

Patients with postoperative initiation of LMWHthromboprophylaxis may safely undergo single-in-jection and continuous catheter techniques. Man-agement is based on total daily dose, timing of thefirst postoperative dose, and dosing schedule.

Twice Daily Dosing. This dosage regimenmay be associated with an increased risk of spinalhematoma. The first dose of LMWH should be ad-ministered no earlier than 24 hours postopera-tively, regardless of anesthetic technique, and onlyin the presence of adequate (surgical) hemostasis.Indwelling catheters should be removed prior toinitiation of LMWH thromboprophylaxis. If a con-tinuous technique is selected, the epidural cathetermay be left indwelling overnight and removed thefollowing day, with the first dose of LMWH admin-istered 2 hours after catheter removal.


Single Daily Dosing. This dosing regimen ap-proximates the European application. The firstpostoperative LMWH dose should be administered6 to 8 hours postoperatively. The second postoper-ative dose should occur no sooner than 24 hoursafter the first dose. Indwelling neuraxial cathetersmay be safely maintained. However, the cathetershould be removed a minimum of 10 to 12 hoursafter the last dose of LMWH. Subsequent LMWHdosing should occur a minimum of 2 hours aftercatheter removal.

Oral Anticoagulants (Warfarin)

Warfarin Pharmacology

Oral anticoagulants, including warfarin, exerttheir anticoagulant effect indirectly by interferingwith the synthesis of the vitamin K-dependent clot-ting factors, factor II (thrombin), VII, IX, and X. Theeffects of warfarin are not apparent until a signifi-cant amount of biologically inactive factors are syn-thesized and is dependent on factor half-life86: fac-tor VII, 6 to 8 hours; factor IX, 24 hours; factor X, 25to 60 hours; and factor II, 50 to 80 hours.

An understanding of the correlation between thevarious vitamin K–dependent factor levels and theINR is critical to regional anesthetic management.Clinical experience with patients who congenitallyare deficient in factors II, IX, or X suggests that afactor activity level of 40% for each factor is ade-quate for normal or near-normal hemostasis.87

Bleeding may occur if the level of any clotting factoris decreased to 20% to 40% of baseline. The PT andINR are most sensitive to the activities of factors VIIand X and are relatively insensitive to factor II.88

Because factor VII has a relatively short half-life,prolongation of the PT and INR may occur in 24 to36 hours. Prolongation of the INR (INR � 1.2)occurs when factor VII activity is reduced to ap-proximately 55% of baseline, while an INR � 1.5 isassociated with a factor VII activity of 40%.86 Thus,an INR �1.5 should be associated with normal he-mostasis.

The same principles apply during recovery of nor-mal hemostasis upon discontinuation of warfarintherapy. Factor VII activity will rapidly increase, asdemonstrated by a decrease in the INR. However,factor II and X activities recover much more slowly;hemostasis may not be adequate even though theINR is 1.4 or less.88 Adequate levels of all vitaminK-dependent factors are typically present when theINR is in the normal range. In emergent situations,the effects of warfarin may be reversed by injectionof vitamin K and/or transfusion of fresh frozenplasma.88

Factors Affecting Warfarin Response

The measured response to anticoagulant therapyat the initiation of treatment varies significantly.Some of the variability may be attributed to druginteractions, but in addition there are patient vari-ables, such as age, female gender, and preexistingmedical conditions (lower patient weight, liver, car-diac, and renal disease) that are associated with anenhanced response to warfarin.89,90 Oriental pa-tients require lower doses than Caucasian patientsduring chronic therapy.89 In addition, there aremany drug interactions described with warfarintherapy that potentiate the anticoagulant effect, in-cluding concomitant administration of antiplateletmedications, heparin, and LMWH.91,92

Warfarin is a drug with a narrow therapeuticrange. Attention to the individual patient’s re-sponse to warfarin therapy and maintenance of aconsistent level of anticoagulation is paramount.Most medical laboratories have a method of con-tacting the caregiver in the event of an excessivelyprolonged PT/INR. However, further precautionsmay be warranted. Inclusion of pharmacy person-nel may be one technique to add consistency inwarfarin management. Because all warfarin ordersare filled by the pharmacy (and entered into acentral computer), linking the pharmacy and labo-ratory results’ computers will allow identification ofpatients with (1) a significant increase in the INR ina predefined time, (2) a subtherapuetic INR, and (3)warfarin therapy without INR assessment. Thepharmacy then notifies the primary service and/orpain service so that appropriate action may betaken. To maintain the desired anticoagulant effect,the patient is instructed in a “warfarin” diet thatcontains foods with a consistent (low) level of vita-min K. These procedures have been successfullyimplemented at the Mayo Clinic.

Neuraxial Techniques in the ChronicallyAnticoagulated Patient

Although no studies have directly examined therisk of procedure-related bleeding and the INR inpatients recently discontinued from warfarin, carefulconsideration should be given before performingneuraxial blocks in these patients. Labeling of warfa-rin in the United States specifically lists spinal punc-ture and lumbar block anesthesia as contraindicatedduring warfarin therapy that is not interrupted priorto surgery.93 Wille-Jorgensen et al.90 reported a caseof difficult epidural placement in a patient fully anti-coagulated with phenprocoumon. The anticoagulanttherapy was unknown to the anesthesiologist. Therewas no bleeding observed during catheter placement,although placement was technically difficult. Satisfac-


tory anesthesia developed and apparently resolved.Three days after surgery, the patient developed paresisof the lower extremities and impairment of the rectaland bladder sphincters. An epidural hematoma wasevacuated from T11-L1, but the extremity paresis wasnot reversed.

Timing of Neuraxial Catheter Removal DuringWarfarin Thromboprophylaxis

The management of patients requiring chronicanticoagulation (with recent discontinuation ofwarfarin in anticipation of surgery) and patientsreceiving warfarin perioperatively for thrombo-prophylaxis remains controversial. Adjusted-dosewarfarin is the most common agent used forthromboembolism prophylaxis after hip and kneereplacement surgery (Table 1). Few data exist re-garding the risk of spinal hematoma in patientswith indwelling spinal or epidural catheters whoare subsequently anticoagulated with warfarin.Bleeding may occur during catheter removal of theepidural catheter as a result of vascular traumaduring catheter manipulation94 or dislodgment ofan existing clot.95

Several studies have examined the use of regionalanesthesia and analgesia in patients who receivedwarfarin during the perioperative period for throm-boembolic prophylaxis. No spinal hematomas werereported in any of the studies; however, the power ofthese studies to detect a rare complication is low.Odoom and Sih96 performed 1,000 continuous lum-bar epidural anesthetics in 950 patients undergoingvascular procedures who were receiving oral antico-agulants preoperatively. The thrombotest (a test mea-suring factor IX activity) was decreased and the aPTTwas prolonged in all patients prior to needle place-ment. A modest heparin infusion was administeredintraoperatively. Epidural catheters remained in placefor 48 hours postoperatively; the coagulation status attime of catheter removal was not described. Therewere no neurologic complications. While the resultsof this study are reassuring, the obsolescence of thethrombotest as a measure of anticoagulation combinedwith the unknown coagulation status of the patientsat the time of catheter removal limit their usefulness.

The use of an indwelling epidural or intrathecalcatheter and the timing of its removal in an antico-agulated patient are also controversial. Althoughthe trauma of needle placement occurs with bothsingle dose and continuous catheter techniques, thepresence of an indwelling catheter could theoreti-cally provoke additional injury to tissue and vascu-lar structures. A combined series of 651 patientsreported no spinal hematomas in patients receivingneuraxial block in conjunction with low-dose war-

farin therapy. The mean INR at the time of catheterremoval was 1.4. However, marked variability inpatient response to warfarin was noted.97,98

There are 2 case reports in the literature describ-ing spinal hematoma in patients who received peri-operative warfarin for thromboembolic prophylaxisand regional anesthesia. Woolson et al.99 reportedan 85-year-old woman who underwent total kneearthroplasty (TKA) with epidural anesthesia andanalgesia. The patient was given a single preopera-tive dose of 10 mg warfarin. Her epidural catheterwas removed on the second postoperative day, atwhich time her INR was 6.3. She developed para-paresis of the lower extremities, which requiredlaminectomy. Badenhorst100 described a female pa-tient who underwent bilateral TKA with epiduralanesthesia and analgesia. This patient also receiveda preoperative dose of warfarin that was continuedthroughout the perioperative period. Her PT themorning of surgery was 14.3 seconds (normal, 11.2to 14.4 seconds; INR not reported). On the thirdpostoperative day, the epidural catheter was re-moved when her PT was 17.3 seconds. At that timeshe complained of blurred vision and tingling andweakness in her right leg. On postoperative day 4,she had bilateral lower extremity sensory and mo-tor deficits. She underwent emergent decompres-sive laminectomy with near complete recovery.Two cases of spinal hematomas in patients antico-agulated with warfarin have been reported throughthe MedWatch system since 1998. The details arescant for both cases. The first patient was chroni-cally anticoagulated with warfarin with a PT greaterthan 50 seconds at the time of needle placement.The second patient received epidural analgesia anddeveloped neurologic deficits 48 hours later (withthe catheter indwelling), at which time her INR was1.6. Although a laminectomy was performed, neu-rologic outcome was not noted.

Regional Anesthetic Management of the Patienton Oral Anticoagulants

The management of patients receiving warfarinperioperatively remains controversial. Consensusstatements are based on warfarin pharmacology,the clinical relevance of vitamin K coagulation fac-tor levels/deficiencies, and the case reports of spinalhematoma among these patients.

Caution should be used when performingneuraxial techniques in patients recently discontin-ued from chronic warfarin therapy. The anticoagu-lant therapy must be stopped, (ideally 4 to 5 daysprior to the planned procedure) and the PT/INRmeasured prior to initiation of neuraxial block.Early after discontinuation of warfarin therapy, the


PT/INR reflect predominantly factor VII levels, anddespite acceptable factor VII levels, factors II and Xlevels may not be adequate for normal hemostasis.Adequate levels of II, VII, IX, and X may not bepresent until the PT/INT is within normal limits.

The concurrent use of medications that affectother components of the clotting mechanisms mayincrease the risk of bleeding complications for pa-tients receiving oral anticoagulants, and do so with-out influencing the PT/INR. These medicationsinclude aspirin and other nonsteroidal anti-inflam-matory drugs (NSAIDs), ticlopidine and clopidogrel,unfractionated heparin, and LMWH.

For patients receiving an initial dose of warfarinprior to surgery, the PT/INR should be checkedprior to neuraxial block if the first dose was givenmore than 24 hours earlier or a second dose of oralanticoagulant has been administered.

Patients receiving low-dose warfarin therapyduring epidural analgesia should have their PT/INRmonitored on a daily basis and checked before cath-eter removal, if initial doses of warfarin are admin-istered more than 36 hours preoperatively. Initialstudies evaluating the safety of epidural analgesia inassociation with oral anticoagulation utilized meandaily doses of approximately 5 mg warfarin. Higherdose warfarin may require more intensive monitor-ing of the coagulation status.

As thromboprophylaxis with warfarin is initiated,neuraxial catheters should be removed when theINR is �1.5. This value was derived from studiescorrelating hemostasis with clotting factor activitylevels greater than 40%.

Neurologic testing of sensory and motor functionshould be performed routinely during epidural an-algesia for patients on warfarin therapy. The type ofanalgesic solution should be tailored to minimizethe degree of sensory and motor block. Thesechecks should be continued after catheter removalfor at least 24 hours, and longer if the INR wasgreater than 1.5 at the time of catheter removal.

An INR � 3 should prompt the physician towithhold or reduce the warfarin dose in patients withindwelling neuraxial catheters. We can make no de-finitive recommendation for removal of neuraxialcatheters in patients with therapeutic levels of antico-agulation during neuraxial catheter infusion.

Reduced doses of warfarin should be given topatients who are likely to have an enhanced re-sponse to the drug.

Antiplatelet Medications

Pharmacology of “Antiplatelet” Medications

Antiplatelet agents include NSAIDs, thienopyri-dine derivatives (ticlopidine and clopidogrel), and

platelet GP IIb/IIIa receptor antagonists (abciximab,eptifibatide, and tirofiban). It is important to notethe pharmacologic differences among the drugswith antiplatelet effects.

Cyclooxygenase (COX) exists in 2 forms. COX-1regulates constitutive mechanisms, while COX-2mediates pain and inflammation. NSAIDs inhibitplatelet COX and prevent the synthesis of throm-boxane A2. Platelets from patients who have beentaking these medications have normal platelet ad-herence to subendothelium and normal primaryhemostatic plug formation. Depending on the doseadministered, aspirin (and other NSAIDs) may pro-duce opposing effects on the hemostatic mecha-nism. For example, platelet COX is inhibited bylow-dose aspirin (60 to 325 mg/d) while largerdoses (1.5 to 2 g/d) will also inhibit the productionof prostacyclin (a potent vasodilator and plateletaggregation inhibitor) by vascular endothelial cells.It has been suggested that the Ivy bleeding time isthe most reliable predictor of abnormal bleeding inpatients receiving antiplatelet drugs. However,there is no evidence to suggest that a bleeding timecan predict hemostatic compromise.101 Plateletfunction is affected for the life of the platelet fol-lowing aspirin ingestion; other nonsteroidal analge-sics (naproxen, piroxicam, ibuprofen) produce ashort-term defect, which normalizes within 3days.102

Celecoxib (Celebrex, Pfizer, New York, NY) andRofecoxib (Vioxx, Merck and Co. Inc., WhitehouseStation, NY) are anti-inflammatory agents that pri-marily inhibit COX-2, an inducible enzyme which isnot expressed in platelets, and thus does not causeplatelet dysfunction.103 After single and multidos-ing, there have not been findings of significantdisruption of platelet aggregation, nor is there ahistory of undesirable bleeding events. The con-comitant use of COX-2 inhibitors and warfarin mayincrease the risk of hemorrhagic complications byincreasing the PT.

The antiplatelet effect of the thienopyridine de-rivatives, ticlopidine and clopidogrel, results frominhibition of adenosine diphosphate (ADP)-inducedplatelet aggregation. These antiplatelet agents, usedin the prevention of cerebrovascular thromboem-bolic events, affect both primary and secondaryplatelet aggregation. Ticlopidine (Ticlid, Roche Lab-oratories, Nutley, NJ) and clopidogrel (Plavix, Bris-tol-Myers Squibb Co., New York, NY) also interferewith platelet-fibrinogen binding and subsequentplatelet-platelet interactions.104 Thienopyridine de-rivatives demonstrate both time- and dose-depen-dent effects; steady state is achieved within 7 daysfor clopidogrel and 14 to 21 days for ticlopidine.Although often administered in combination with


aspirin, the safety of concomitant use of clopi-dogrel or ticlopidine and aspirin has not beenestablished and may be associated with increasedrisk of hemorrhagic events. Serious hematologicadverse reactions, including agranulocytosis,thrombotic thrombocytopenic purpura, andaplastic anemia have resulted in placement of ablack box warning on ticlopidine. Labeling rec-ommends, “If a patient is to undergo electivesurgery, and an antiplatelet effect is not desired,clopidogrel should be discontinued 7 days andticlopidine 10-14 days, prior to surgery.”

Platelet GP IIb/IIIa receptor antagonists, includ-ing abciximab (Reopro, Eli Lilly and Co., Indianap-olis, IN), eptifibatide (Integrilin, Millenium Phar-maceuticals Inc., San Francisco, CA), and tirofiban(Aggrastat, Merck and Co.), inhibit platelet aggre-gation by interfering with platelet-fibrinogen andplatelet-von Willebrand factor binding. Because fi-brinogen and von Willebrand factor have multiplebinding sites, they can bind to multiple platelets,causing cross-linking and platelet aggregation. Con-versely, inhibition of GP IIb/IIIa receptors blocksthe final common pathway to platelet aggregation.The majority of clinical trials involving the GP IIb/IIIa antagonists have evaluated their use in thetreatment of acute coronary syndrome (with orwithout percutaneous coronary intervention). Im-portantly, the GP IIb/IIIa antagonists are typicallyadministered in combination with aspirin and hep-arin. Contraindications include a history of surgerywithin 4 to 6 weeks. Time to normal platelet aggre-gation following discontinuation of therapy rangesfrom 8 hours (eptifibatide, tirofiban) to 24 to 48hours (abciximab). During therapy with GP IIb/IIIaantagonists, labeling precautions recommend thatpuncture of noncompressible sites and “epidural”procedures be avoided.

Spinal Hematoma in Patients ReceivingAntiplatelet Medications

At the 1998 Consensus Conference on NeuraxialAnesthesia and Anticoagulation, it was concludedthat NSAIDs, in and of themselves, did not appearto present significant risk to patients for developingspinal-epidural hematomas.11 Vandermeulen etal.26 implicated antiplatelet therapy in 3 of the 61cases of spinal hematoma occurring after spinal orepidural anesthesia. These patients had receivedaspirin, indomethacin, or ticlopidine. Three addi-tional case reports related to neuraxial techniqueshave been published in recent years, 2 involvingketorolac and 2 involving a thienopyridine deriva-tive.105-108 The paucity of case reports is important,given the prevalence of NSAID use among the gen-

eral population and that subset of patients withacute, chronic, and/or cancer pain-related problemswho subsequently receive interventional therapy.

Several large studies have demonstrated the rel-ative safety of central neural block in combinationwith antiplatelet therapy, although the total num-ber of patients in this combined series is only4,714.11 If low-dose aspirin creates the greatest im-pact on platelet function, patients receiving 60 to325 mg aspirin would theoretically represent thegreatest risk of significant bleeding. However, theCollaborative Low-dose Aspirin Study in Pregnancy(CLASP) Group included 1,422 high-risk obstetricpatients administered 60 mg aspirin daily who un-derwent epidural anesthesia without any neuro-logic sequelae.109 A recent prospective study eval-uated the risk of neurologic complications followingepidural steroid injection. There were no spinal he-matomas among the 1,214 patients, including the32% of patients who reported NSAID use prior toinjection.110 These results confirm those of previousstudies performed in obstetric and surgical popula-tions.

No series involving the performance of neuraxialblock in the presence of thienopyridine derivativesor platelet GP IIb/IIIa receptor antagonists has beenperformed. Although the data are inconsistent, in-creased perioperative bleeding in patients undergo-ing cardiac and vascular surgery after receivingticlopidine, clopidogrel, and GP IIb/IIIa antagonistshas been noted.111,112 In general, the cardiac surgi-cal112 and interventional radiology literature rec-ommend that elective surgery be delayed 24 to 48hours following abciximab and 4 to 8 hours follow-ing eptifibatide or tirofiban.113 Surgery performedwithin 12 hours of abciximab administration willmost likely necessitate a platelet transfusion. Therehave been 3 spinal hematomas attributed toneuraxial techniques and ticlopidine or clopidogrel,including 1 patient undergoing a series of epiduralsteroid injections.106,108,114 Two cases of severebleeding following lumbar sympathetic block in pa-tients on ticlopidine and clopidogrel were recentlyreported, which may warrant concern regardingthe risk of anesthesia-related hemorrhagic compli-cations.16

Combination of Antiplatelet Medications WithAnticoagulants and Thrombolytics

NSAIDs alone do not significantly increase therisk of spinal hematoma. However, combinationtherapy with unfractionated heparin, LMWH, oralanticoagulants, and thrombolytics have been dem-onstrated to increase the frequency of spontaneoushemorrhagic complications, bleeding at puncture


sites, and spinal hematoma.14,18,22,42 For example,in the series of 40 spinal hematomas associated withLMWH reported in 1998, 10 patients received con-comitant antiplatelet medications. Likewise, in acase report of spinal hematoma following epiduralsteroid injection, Benzon et al.106 noted the patienthad received multiple antiplatelet medications, in-cluding clopidogrel and aspirin.

Anesthetic Management of the Patient ReceivingAntiplatelet Medications

Antiplatelet medications, including NSAIDs, thien-opyridine derivatives (ticlopidine and clopidogrel),and platelet GP IIb/IIIa antagonists (abciximab, eptifi-batide, tirofiban) exert diverse effects on platelet func-tion. The pharmacologic differences make it impossi-ble to extrapolate between the groups of drugsregarding the practice of neuraxial techniques.

There is no wholly accepted test, including thebleeding time, which will guide antiplatelet ther-apy. Careful preoperative assessment of the patientto identify alterations of health that might contrib-ute to bleeding is crucial. These conditions include ahistory of easy bruisability/excessive bleeding, fe-male gender, and increased age.

NSAIDs appear to represent no added significantrisk for the development of spinal hematoma inpatients having epidural or spinal anesthesia. Theuse of NSAIDs alone does not create a level of riskthat will interfere with the performance ofneuraxial blocks.

At this time, there do not seem to be specificconcerns as to the timing of single-shot or cathetertechniques in relationship to the dosing of NSAIDs,postoperative monitoring, or the timing ofneuraxial catheter removal.

The actual risk of spinal hematoma with ticlopi-dine and clopidogrel and the GP IIb/IIIa antagonists

is unknown. Consensus management is based onlabeling precautions and the surgical, interven-tional cardiology/radiology experience.

Based on labeling and surgical reviews, the sug-gested time interval between discontinuation ofthienopyridine therapy and neuraxial block is 14days for ticlopidine and 7 days for clopidogrel.

Platelet GP IIb/IIIa inhibitors exert a profoundeffect on platelet aggregation. Following adminis-tration, the time to normal platelet aggregation is24 to 48 hours for abciximab and 4 to 8 hours foreptifibatide and tirofiban. Neuraxial techniquesshould be avoided until platelet function has recov-ered. Although GP IIb/IIIa antagonists are contra-indicated within 4 weeks of surgery, should one beadministered in the postoperative period (followinga neuraxial technique), the patient should be care-fully monitored neurologically.

The concurrent use of other medications affectingclotting mechanisms, such as oral anticoagulants,unfractionated heparin, and LMWH, may increasethe risk of bleeding complications. COX-2 inhibitorshave minimal effect on platelet function and shouldbe considered in patients who require anti-inflam-matory therapy in the presence of anticoagulation.

Effects of Herbal Therapies onCoagulation

There is a widespread use of herbal medicationsin surgical patients. Most patients do not volunteerinformation regarding herbal medication use; ob-taining such a history may be difficult.115-117 Mor-bidity and mortality associated with herbal use maybe more likely in the perioperative period becauseof the polypharmacy and physiological alterationsthat occur. Such complications include bleedingfrom garlic, ginkgo, and ginseng, and potential in-teraction between ginseng-warfarin (Table 5). Be-

Table 5. Three Herbal Medications With the Greatest Impact on Hemostasis

Herb Important Effects Perioperative Concerns

Time to NormalHemostasis AfterDiscontinuation

Garlic Inhibition of platelet aggregation (may beirreversible)

Potential to increase bleeding, especially whencombined with other medications that inhibitplatelet aggregation

7 days

Increased fibrinolysisEquivocal antihypertensive activity

Ginko Inhibition of platelet-activating factor Potential to increase bleeding, especially whencombined with other medications that inhibitplatelet aggregation

36 hours

Ginseng Lowers blood glucose Hypoglycemia 24 hoursIncreased prothrombin and activated

partial prothrombin times in animalsPotential to increase risk of bleeding

Other diverse effects Potential to decrease anticoagulant effect ofwarfarin

NOTE. At this time, it is not deemed necessary to discontinue herbal medications and allow resolution of their effects on hemostasisprior to surgery or anesthesia.


cause the current regulatory mechanism for com-mercial herbal preparations sold in the UnitedStates does not adequately protect against unpre-dictable or undesirable pharmacological effects, it isespecially important for anesthesiologists to be fa-miliar with related literature on herbal medicationswhen caring for patients in the perioperative pe-riod. Sources for reliable and updated informationare important in helping anesthesiologists stayabreast of new discoveries about the effects ofherbal medications in humans. Several resourcesare available on the World Wide Web as clinicalaids. The Center for Food Safety and Applied Nu-trition, Food and Drug Administration (http://vm.cfsan.fda.gov/�dms/supplmnt.html) and NationalCenter for Complementary and Alternative Medi-cine, National Institutes of Health (http://nccam.nih.gov) websites contain fact sheets about alternativetherapies, consensus reports, and databases. TheFDA website may also be used to report adverseevents.

Garlic

Garlic is one of the most extensively researchedmedicinal plants. It has the potential to modify therisk of developing atherosclerosis by reducing bloodpressure, thrombus formation, and serum lipid andcholesterol levels.118 The usual dosage is 4 g (�2cloves) of fresh bulb or its equivalent as an extractor tincture per day.119 Garlic inhibits in vivo plateletaggregation in a dose-dependent fashion. The effectof one of its constituents, ajoene, appears to beirreversible and may potentiate the effect of otherplatelet inhibitors, such as prostacyclin, forskolin,indomethacin, and dipyridamole.120,121 Althoughthese effects have not been consistently demon-strated in volunteers, there is 1 case in the literatureof an octagenarian who developed a spontaneousepidural hematoma that was attributed to heavygarlic use.122

Ginkgo

Ginkgo is derived from the leaf of Ginkgo biloba.It has been used in cognitive disorders, peripheralvascular disease, age-related macular degeneration,vertigo, tinnitus, erectile dysfunction, and altitudesickness.123,124 The compounds believed to be re-sponsible for its pharmacologic effects are the ter-penoids and flavonoids. The usual dosage is 120 to240 mg standardized extract per day in 2 or 3 di-vided doses.125 Ginkgo appears to inhibit platelet-activating factor (PAF).126 Clinical trials in a smallnumber of patients have not demonstrated bleedingcomplications, but 4 reported cases of spontaneous

intracranial bleeding127-130 have been associatedwith ginkgo use.

Ginseng

Among the several species used for pharmaco-logic effects, Asian ginseng and American ginsengare the most commonly described. Ginseng hasbeen labeled an “adaptogen” because it reputedlyprotects the body against stress and restores ho-meostasis.131 The usual dosage is 1 to 2 g root or 200mg standardized extract per day.132 Ginseng has abroad, but incomplete, pharmacologic profile, be-cause it has many heterogeneous and sometimesopposing effects of different ginsenosides.133 Thereis a concern about ginseng’s effect on coagulationpathways. Ginsenosides inhibit platelet aggregationin vitro134,135 and prolong both thrombin time andactivated partial thromboplastin time in rats.136

These findings await confirmation in humans. Al-though ginseng may inhibit the coagulation cas-cade, ginseng use was associated with a significantdecrease in warfarin anticoagulation in 1 reportedcase.137

Anesthetic Management of the Patient ReceivingHerbal Therapy

Herbal drugs, by themselves, appear to representno added significant risk for the development ofspinal hematoma in patients having epidural orspinal anesthesia. This is an important observation,because it is likely that a significant number of oursurgical patients utilize alternative medications pre-operatively and perhaps during their postoperativecourse.

The use of herbal medications alone does notcreate a level of risk that will interfere with theperformance of neuraxial blocks. Mandatory dis-continuation of these medications or cancellation ofsurgery in patients in whom these medicationshave been continued is not supported by availabledata.

Data on the combination of herbal therapy withother forms of anticoagulation are lacking. How-ever, the concurrent use of other medications af-fecting clotting mechanisms, such as oral anticoagu-lants or heparin, may increase the risk of bleedingcomplications in these patients.

There is no wholly accepted test to assess ade-quacy of hemostasis in the patient reporting preop-erative herbal medications.

At this time, there do not seem to be specificconcerns as to the timing of neuraxial block inrelationship to the dosing of herbal therapy, post-operative monitoring, or the timing of neuraxialcatheter removal.


New Anticoagulants

New antithrombotic drugs, which target varioussteps in the hemostatic system, such as inhibitingplatelet aggregation, blocking coagulation factors,or enhancing fibrinolysis, are continually underdevelopment. The most extensively studied are an-tagonists of specific platelet receptors and directthrombin inhibitors. Many of these antithromboticagents have prolonged half-lives and are difficult toreverse without administration of blood compo-nents. It is likely that orally bioavailable “heparins”will be introduced in the near future. The adminis-tration of these medications in combination withneuraxial anesthesia must be carefully considered.

Thrombin Inhibitors

Recombinant hirudin derivatives, including desiru-din (Revasc, Aventis Pharmaceuticals, Paris, France),lepirudin (Refludan, Aventis Pharmaceuticals), andbivalirudin (Angiomax, The Medicine Co., Cam-bridge, MA) inhibit both free and clot-bound throm-bin. Argatroban (Acova), an L-arginine derivative, hasa similar mechanism of action. These medications areindicated for the treatment and prevention of throm-bosis in patients with heparin-induced thrombocyto-penia and as an adjunct to angioplasty proce-dures.138,139 Desirudin has also been evaluated forprevention of DVT/PE following hip replace-ment.140 The anticoagulant effect of thrombin in-hibitors is monitored by the aPTT, and is present for1 to 3 hours after IV administration. Hemorrhagiccomplications, particularly when combined withthrombolytic or antiplatelet agents, may be lifethreatening. There is no “antidote”; the antithrom-bin effect cannot be reversed pharmacologically.Although there are no case reports of spinal hema-toma related to neuraxial anesthesia among pa-tients who have received a thrombin inhibitor,spontaneous intracranial bleeding has been re-ported. Due to the lack of information available, nostatement regarding risk assessment and patientmanagement can be made. Identification of cardi-ology and surgical risk factors associated with bleed-ing following invasive procedures may be helpful.

Fondaparinux

Fondaparinux, an injectable synthetic pentasac-charide, was approved in December 2001. The FDAreleased fondaparinux (Arixtra, Sanofi-Synthelabo,West Orange, NJ) with a black box warning similarto that of the LMWHs and heparinoids. Fondapa-rinux produces its antithrombotic effect throughfactor Xa inhibition. The plasma half-life offondaparinux is 21 hours, allowing for single daily

dosing, with the first dose administered 6 hourspostoperatively.141 Investigators reported a spinalhematoma among the initial dose-ranging study (ata dose that was subsequently determined to betwice required for thromboprophylaxis).142,143 Noadditional spinal hematomas were reported in thecombined series of 3,600 patients who underwentspinal or epidural anesthesia in combination withfondaparinux thromboprophylaxis. However, theconditions for performance of neuraxial block werestrictly controlled. Patients were included in subse-quent clinical trials only if needle placement wasatraumatic and accomplished on the first attempt.In addition, indwelling epidural catheters were re-moved 2 hours prior to fondaparinux administra-tion.143 These practice guidelines may not be feasi-ble in clinical practice. For example, in a prospectiveseries, less than 40% of neuraxial blocks were suc-cessful with one pass.24

Anesthetic Management of the Patient ReceivingFondaparinux

The actual risk of spinal hematoma with fondapa-rinux is unknown. Consensus statements are basedon the sustained and irreversible antithromboticeffect, early postoperative dosing, and the spinalhematoma reported during initial clinical trials.Close monitoring of the surgical literature for riskfactors associated with surgical bleeding may behelpful in risk assessment and patient management.

Until further clinical experience is available, per-formance of neuraxial techniques should occur un-der conditions utilized in clinical trials (single nee-dle pass, atraumatic needle placement, avoidance ofindwelling neuraxial catheters). If this is not feasi-ble, an alternate method of prophylaxis should beconsidered.

German and Spanish Guidelines forThromboembolism Prophylaxis andRegional Anesthesia

Every year, several million neuraxial blocks areperformed in Europe, the majority on patients un-dergoing in-patient surgery who receive thrombo-embolic prophylaxis, usually with either unfrac-tionated or LMWH. The reported incidence ofclinically important spinal bleeding resulting in per-manent neurologic lesions is extremely low. It isimportant to note that 70% to 75% of neuraxialblocks performed in Europe are single-dose spinalanesthetics; continuous epidural techniques ac-count for only 19% of central blocks.69

Consensus statements tend to reflect the clinicalexperience and concerns of the conference partici-


pants. At least 2 other societies have approved of-ficial guidelines for thromboembolism prophylaxisand regional anesthesia. A comparison of the Span-ish Consensus Statements76 with those of the Ger-man Society of Anesthesiology and Intensive Care77

and the ASRA reveals several similarities as well asdifferences (Table 6). The management of patientsreceiving unfractionated heparin is remarkably sim-ilar. As expected, the ASRA guidelines for LMWHare much more conservative than the correspond-ing European statements owing to the large num-ber of hematomas in North America. It is notablethat the Spanish Consensus Forum extensively an-alyzes the risk of spinal hematoma associated withall antiplatelet agents and recommends discontinu-ation of these medications prior to neuraxial block.It is somewhat difficult to justify this practice; thesafety of neuraxial anesthesia in combination withantiplatelet therapy is well documented11 and only4 spinal hematomas in patients on NSAIDs havebeen reported.26,105,107 In addition, nearly half ofpatients report preoperative antiplatelet therapywhen questioned.24 If a significant risk existed, onewould expect more published reports of spinal he-matoma among these patients. Conversely, morepotent antiplatelet agents, including ticlopidine,clopidogrel, and abciximab, warrant increased cau-

tion. Revised ASRA guidelines recommend thatneuraxial blocks be avoided in patients receivingthese medications until platelet function has recov-ered. It is of interest that neither the Spanish northe German Consensus Statements address the is-sue of neuraxial anesthesia in the patient receivingthrombolytics or herbal medications. Admittedly,although the pharmacology of thrombolytic agentsis fairly well defined, few data are available to quan-tify the risk. Additional information is needed toallow evidence-based recommendations for pa-tients reporting herbal therapy.

Unplanned Anticoagulation DuringNeuraxial Analgesia

Occasionally, patients require emergent anti-thrombotic therapy (vascular graft thrombosis,acute coronary syndrome/myocardial infarction),or a breakdown in communication results in unan-ticipated anticoagulation in the presence of indwell-ing epidural catheters. It is critical that the PainService be aware of alterations in the degree andtiming of anticoagulation. Increasing centralizationand computerization make it possible for HospitalPharmacy Services to assist with patient manage-ment. Because all medication orders are filled by

Table 6. Neuraxial Anesthesia in the Patient Receiving Thromboprophylaxis

AntiplateletMedications

Unfractionated Heparin

Warfarin Thrombolytics Herbal TherapySubcutaneous IntravenousLow MolecularWeight Heparin

German Society ofAnesthesiologyand Intensive CareMedicine

No contraindication Needle placement4 h after heparin;heparin 1 h afterneedle placementor catheterremoval

Needle placementand/or catheterremoval 4 h afterdiscontinuingheparin, heparinize1 h after neuraxialtechnique; delaysurgery 12 h iftraumatic

Neuraxial technique10-12 h afterLMWH; next dose 4h after needle orcatheter placement

Discontinue inadvance, removecatheter prior toinitiation ofwarfarin

Not discussed Not discussed

SpanishConsensus Forum

Discontinue inadvance

Not discussed Neuraxialtechnique 4 hafter heparindose; heparinize30 min afterneedle placement;delayheparinization 6 hif traumatic

Needle placement12� h after LMWH;first postoperativedose 4-12 h;catheters removed10-12 h afterLMWH and 4 hprior to next dose;postpone LMWH 24if traumatic

INR � 1.5 forperformance ofneuraxialtechniques; noINR guidelinesfor catheterremoval

Not discussed Not discussed

American Societyof RegionalAnesthesia andPain Medicine

No contraindicationwith NSAIDs;discontinueticlopidine 14 d,clopidogrel 7 d, GPIIb/IIIa inhibitors8-48 h in advance

No contraindication,consider delayingheparin until afterblock if technicaldifficulty anticipated

Heparinize 1 hafter neuraxialtechnique,remove catheter2-4 h after lastheparin dose; nomandatory delayif traumatic

Twice daily dosing:LMWH 24 h aftersurgery, regardlessof technique;remove neuraxialcatheter 2 h beforefirst LMWH dose.Single daily dosing:according toEuropeanstatements

Document normalINR afterdiscontinuation(prior to neuraxialtechnique);remove catheterwhen INR � 1.5(initiation oftherapy)

No data onsafety intervalforperformanceof neuraxialtechnique orcatheterremoval;followfibrinogenlevel

No evidence formandatorydiscontinuationprior to neuraxialtechnique; beaware ofpotential druginteractions

Abbreviations: NSAIDs, nonsteroidal anti-inflammatory drugs; GP IIb/IIIa, platelet glycoprotein receptor IIb/IIIa inhibitors; INR, international normalizedratio; LMWH, low molecular weight heparin.

Data from the German Society of Anesthesiology and Intensive Care Medicine Consensus guidelines77 and the Spanish Consensus Forum.76


pharmacists using a central computer, patients whoreceive an epidural infusion are identified withinthe pharmacy database. Any subsequent order foran antithrombotic agent is flagged as a drug “inter-action” during entry, and the pharmacist receivesan alert notice to contact the Pain Service. This“pharmacy failsafe” allows the Pain Service to par-ticipate proactively in the timing of catheter re-moval and subsequent anticoagulation, as well asclosely monitor the patient’s neurologic status.144

Plexus and Peripheral Block in theAnticoagulated Patient

Although spinal hematoma is the most significanthemorrhagic complication of regional anesthesiadue to the catastrophic nature of bleeding into afixed and noncompressible space, the associatedrisk following plexus and peripheral techniques re-mains undefined. There are no investigations thatexamine the frequency and severity of hemorrhagiccomplications following plexus or peripheral blockin anticoagulated patients. However, few reportsof serious complications following neurovascularsheath cannulation for surgical, radiological, or car-diac indications have been reported. For example,during interventional cardiac procedures, large borecatheters are placed within brachial or femoral ves-sels. Heparin, LMWH, antiplatelet medications,and/or thrombolytics are subsequently adminis-tered. In a series of 4,879 patients undergoing car-diac catheterization and/or coronary angioplasty,the frequency of vascular complications was 0.39%.Size of the catheter (5-French v 7-French) and de-gree of anticoagulation influenced the frequency ofcomplications.145 No neurologic complications oc-curred as a result of the vascular injury; 1 patientrequired transfusion.

Several cases of vascular injury with (or without)resultant nerve dysfunction have been describedfollowing plexus or peripheral techniques in pa-tients with normal146-149 and abnormal15,16,150,151

hemostasis. In all patients with neurodefi-cits,147,149,150 neurologic recovery was completewithin 6 to 12 months. Thus, while bleeding into aneurovascular sheath may result in significant de-creases in hematocrit, the expandable nature ofperipheral site may decrease the chance of irrevers-ible neural ischemia.

All cases of major bleeding associated with non-neuraxial techniques occurred after psoas compart-ment or lumbar sympathetic block. One patientwith a mechanical heart valve had been therapeu-tically anticoagulated with warfarin and was con-verted to standard heparin in anticipation of kneearthroscopy. Coagulation status was normal at the

time of psoas compartment block. However, a hep-arin infusion (and oral warfarin) was restarted 8hours after needle placement. On the third postop-erative day, the patient was noted to have markedlyprolonged aPTT and elevated PT. CT performed toevaluate the patient’s complaint of flank pain andongoing reduction in hemoglobin (from 12.8 g/dLto 8.8 g/dL) noted a large hematoma arising fromthe psoas muscle. Vitamin K and 2 U blood wereadministered.15

Two of the cases associated with psoas compart-ment block received LMWH perioperatively. In 1case, a continuous psoas block and single injectionsciatic block were performed prior to knee arthro-plasty. Although frank blood was noted in the psoascatheter, after partial withdrawal, aspiration andtest dose were negative. Enoxaparin was adminis-tered 40 hours postoperatively; the catheter wasremoved 2 hours after LMWH injection, duringpeak anticoagulant effect. The patient complainedof severe paravertebral pain 4 hours later, but re-mained neurologically intact. A CT scan demon-strated a large retroperitoneal hematoma. Herhemoglobin had decreased from 14.4 g/dL on ad-mission to 7.1 g/dL, necessitating transfusion of 4 Ured blood cells.15 An additional LMWH-related he-matoma occurred in a patient hospitalized with acalcaneal fracture.150 Thromboprophylaxis was ini-tiated with enoxaparin 30 mg twice daily. Surgicaldebridement had been previously performed twiceunder uneventful combined psoas-sciatic block. Ap-proximately 20 hours after a dose of enoxaparin,she was brought to the operating room for a thirdprocedure. Attempts at psoas needle placementwere unsuccessful and traumatic. The followingday, the patient complained of hip pain, whichprogressed to weakness. A retroperitoneal hema-toma involving the psoas muscle was noted on CT.She gradually regained all sensory and motor func-tion with conservative therapy.

Thienopyridine derivatives (ticlopidine and clopi-dogrel) were implicated in 2 cases of severe bleed-ing following lumbar sympathetic block.16 The firstpatient underwent 2 lumbar sympathetic blocks.His medications included ticlopidine 500 mg/d. Nocoagulation testing was performed prior to needleinsertion. After the first block, his hemoglobin de-creased from 13.5 g/dL to 10.3 g/dL, and he com-plained of groin pain and medial thigh numbness.Vessel puncture occurred during performance ofthe second block 6 days later. The following night,he complained of severe groin pain, accompaniedby a decrease in his blood pressure and hemoglobin(8.9 g/dL) due to a large retroperitoneal hematoma.The patient was transfused and achieved full recov-ery. The second patient underwent uneventful


lumbar sympathetic block. Her clopidogrel hadbeen discontinued 3 days prior to the block; coag-ulation parameters, including bleeding time werenormal. Nine hours later, she acutely complained ofgroin pain. The patient was found pulseless 3 hourslater; attempts at resuscitation were unsuccessful.Autopsy revealed a 2 to 3 L hematoma in the psoasmuscle.

These cases suggest that significant blood loss,rather than neural deficits, may be the most seriouscomplication of non-neuraxial regional techniquesin the anticoagulated patient. Additional informa-tion is needed to make definitive recommendations.Conservatively, the Consensus Statements onNeuraxial Anesthesia and Anticoagulation may beapplied to plexus and peripheral techniques.15

However, this may be more restrictive than neces-sary.

Practice guidelines or recommendations summa-rize evidence-based reviews. However, the rarity ofspinal hematoma defies a prospective-randomizedstudy, and there is no current laboratory model. Asa result, these consensus statements represent thecollective experience of recognized experts in thefield of neuraxial anesthesia and anticoagulation.They are based on case reports, clinical series, phar-macology, hematology, and risk factors for surgicalbleeding. An understanding of the complexity ofthis issue is essential to patient management; a“cookbook” approach is not appropriate. Rather,the decision to perform spinal or epidural anesthe-sia/analgesia and the timing of catheter removal ina patient receiving antithrombotic therapy shouldbe made on an individual basis, weighing the small,though definite risk of spinal hematoma with thebenefits of regional anesthesia for a specific patient.Alternative anesthetic and analgesic techniques ex-ist for patients considered an unacceptable risk. Thepatient’s coagulation status should be optimized atthe time of spinal or epidural needle/catheter place-ment, and the level of anticoagulation must becarefully monitored during the period of epiduralcatheterization. Indwelling catheters should not beremoved in the presence of therapeutic anticoagu-lation, as this appears to significantly increase therisk of spinal hematoma. It must also be remem-bered that identification of risk factors and estab-lishment of guidelines will not completely eliminatethe complication of spinal hematoma. In Vander-meulen’s series, although 87% of patients had ahemostatic abnormality or difficulty with needlepuncture, 13% had no identifiable risk factor.26

Vigilance in monitoring is critical to allow earlyevaluation of neurologic dysfunction and promptintervention. We must focus not only on the pre-

vention of spinal hematoma, but also optimizationof neurologic outcome.

References

1. Liu S, Carpenter RL, Neal JM. Epidural anesthesiaand analgesia. Their role in postoperative outcome.Anesthesiology 1995;82:1474-1506.

2. Tuman KJ, McCarthy RJ, March RJ, DeLaria GA,Patel RV, Ivankovich AD. Effects of epidural anes-thesia and analgesia on coagulation and outcomeafter major vascular surgery. Anesth Analg 1991;73:696-704.

3. Modig J. The role of lumbar epidural anaesthesia asantithrombotic prophylaxis in total hip replace-ment. Acta Chirurgica Scand 1985;151:589-594.

4. Capdevila X, Barthelet Y, Biboulet P, Ryckwaert Y,Rubenovitch J, d’Athis F. Effects of perioperativeanalgesic technique on the surgical outcome andduration of rehabilitation after major knee surgery.Anesthesiology 1999;91:8-15.

5. American College of Chest Physicians and the Na-tional Heart, Lung and Blood Institute NationalConference on Antithrombotic Therapy. Chest 1986;89:1S-106S.

6. Geerts WH, Heit JA, Clagett GP, Pineo GF, ColwellCW, Anderson FA, Wheeler HB. Prevention of ve-nous thromboembolism. [Review] Chest 2001;119:132S-175S.

7. Horlocker TT, Wedel DJ. Anticoagulation andneuraxial block: Historical perspective, anestheticimplications, and risk management. Reg Anesth PainMed 1998;23:129-134.

8. Baron EM, Burke JA, Akhtar N, Young WF. Spinalepidural hematoma associated with tissue plasmin-ogen activator treatment of acute myocardial infarc-tion. Cathet Cardiovasc Intervent 1999;48:390-396.

9. Shaieb MD, Watson BN, Atkinson RE. Bleedingcomplications with enoxaparin for deep venousthrombosis prophylaxis. J Arthroplasty 1999;14:432-438.

10. Enneking KF, Benzon HT. Oral anticoagulants andregional anesthesia: A perspective. Reg Anesth PainMed 1998:23:140-145.

11. Urmey WF, Rowlingson JC. Do antiplatelet agentscontribute to the development of perioperative spi-nal hematoma? Reg Anesth Pain Med 1998;23:146-151.

12. Rosenquist RW, Brown DL. Neuraxial bleeding: Fi-brinolytics/thrombolytics. Reg Anesth Pain Med 1998;23:152-156.

13. Liu SS, Mulroy MF. Neuraxial anesthesia and anal-gesia in the presence of standard heparin. Reg AnesthPain Med 1998;23:157-163.

14. Horlocker TT, Wedel DJ. Neuraxial block and lowmolecular weight heparin: Balancing perioperativeanalgesia and thromboprophylaxis. Reg Anesth PainMed 1998;23:164-177.

15. Weller RS, Gerancher JC, Crews JC, Wade KL. Ex-tensive retroperitoneal hematoma without neuro-logic deficit in two patients who underwent lumbar


plexus block and were later anticoagulated. Anesthe-siology 2003;98:581-583.

16. Maier C, Gleim M, Weiss T, Stachetzki U, Nicolas V,Zenz M. Severe bleeding following lumbar sympa-thetic blockade in two patients under medicationwith irreversible platelet aggregation inhibitors. An-esthesiology 2002;97:740-743.

17. Fourth American College of Chest Physicians Con-sensus Conference on Antithrombotic Therapy.Chest 1995;108:225S-522S.

18. Murray DW, Britton AR, Bulstrode CJ. Thrombo-prophylaxis and death after total hip replacement.J Bone Joint Surg Br 1996;78:863-870.

19. Mantilla CB, Horlocker TT, Schroeder DR, Berry DJ,Brown DL. Frequency of myocardial infarction, pul-monary embolism, deep venous thrombosis, anddeath following primary hip or knee arthroplasty.Anesthesiology 2002;96:1140-1146.

20. Ryan TJ, Antman EM, Brooks NH, Califf RM, HillisLD, Hiratzka LF, Rapaport E, Riegel B, Russell RO,Smith EE 3rd, Weaver WD, Gibbons RJ, Alpert JS,Eagle KA, Gardner TJ, Garson A Jr, Gregoratos G,Ryan TJ, Smith SC Jr. 1999 update: ACC/AHAguidelines for the management of patients withacute myocardial infarction. A report of the Amer-ican College of Cardiology/American Heart Associ-ation Task Force on Practice Guidelines (Committeeon Management of Acute Myocardial Infarction).J Am Coll Cardiol 1999;34:890-911.

21. Levine MN, Raskob G, Landefeld S, Kearon C. Hem-orrhagic complications of anticoagulant treatment.[Review] Chest 2001;119:108S-121S.

22. Levine MN, Goldhaber SZ, Gore JM, Hirsh J, CaliffRM. Hemorrhagic complications of thrombolytictherapy in the treatment of myocardial infarctionand venous thromboembolism. [Review] Chest1995;108:291S-301S.

23. Tryba M. Ruckmarksnahe regionalanasthesie undniedermolekulare heparine: Pro. Anasth Inten-sivmed Notfallmed Schmerzther 1993;28:179-181.

24. Horlocker TT, Wedel DJ, Schroeder DR, Rose SH,Elliott BA, McGregor DG, Wong GY. Preoperativeantiplatelet therapy does not increase the risk ofspinal hematoma with regional anesthesia. AnesthAnalg 1995;80:303-309.

25. Stafford-Smith M. Impaired haemostasis and re-gional anaesthesia. Can J Anaesth 1996:43:R129-141.

26. Vandermeulen EP, Van Aken H, Vermylen J. Anti-coagulants and spinal-epidural anesthesia. AnesthAnalg 1994;79:1165-1177.

27. Cheney FW, Domino KB, Caplan RA, Posner KL.Nerve injury associated with anesthesia: A closedclaims analysis. Anesthesiology 1999;90:1062-1069.

28. Majerus PW, Broze GJ, Miletich JP, Tollefson PM.Goodman and Gilman’s The Pharmacological Basis ofTherapeutics, ed 8. New York, NY: McGraw-Hill,1993:1322-1331.

29. DePorto R, Ahn DH, Gianutsos JG. Paraplegia sub-sequent to administration of tissue plasminogen ac-tivator and intravenous heparin following myocar-

dial infarction—A case report. J Spinal Cord Med2000;23:150-152.

30. Lopez AG, Lara JMP, Hidalgo RH, Gonzalo EP. Spi-nal epidural hematoma following thrombolytictherapy for acute myocardial infarction. Orthopedics1999;22:987-988.

31. Chan KC, Wu DJ, Ueng KC, Lin CS, Tsai CF, ChenKS, Lin MC, Wang KL, Lin CS. Spinal epiduralhematoma following tissue plasminogen activatorand heparinization for acute myocardial infarction.Jpn Heart J 2002;43:417-421.

32. Groen R J, van Alphen HA. Operative treatment ofspontaneous spinal epidural hematomas: A study ofthe factors determining postoperative outcome.Neurosurgery 1996;39:494-509.

33. Connolly ES Jr, Winfree CJ, McCormick PC. Man-agement of spinal epidural hematoma after tissueplasminogen activator: A case report. Spine 1996;21:1694-1698.

34. Sawin PD, Traynelis VC, Follett KA. Spinal epiduralhematoma following coronary thrombosis with tis-sue plasminogen activator: Report of two cases.J Neurosurg 1995;83:350-353.

35. Wulf H. Epidural anaesthesia and spinal haema-toma. Can J Anaesth 1996;43:1260-1271.

36. Dickman CA, Shedd SA, Spetzler RF, Shetter AG,Sonntag KH. Spinal epidural hematoma associatedwith epidural anesthesia: Complications of systemicheparinization in patients receiving peripheral vas-cular thrombolytic therapy. Anesthesiology 1990;72:947-950.

37. Onishchuk JL, Carlsson C. Epidural hematoma as-sociated with epidural anesthesia: Complications ofanticoagulant therapy. Anesthesiology 1992;77:1221-1223.

38. Rabito SF, Ahmed S, Feinstein L, Winnie AP. Intra-thecal bleeding after the intraoperative use of hep-arin and urokinase during continuous spinal anes-thesia. Anesth Analg 1996;82:409-411.

39. Hirsh J, Raschke R, Warkentin TE, Dalen JE, DeykinD, Poller L. Heparin: Mechanism of action, pharma-cokinetics, dosing considerations, monitoring, effi-cacy, and safety. Chest 1995;108:258S-275S.

40. Murray DJ, Brosnahan WJ, Pennell B, KapalanskiD, Weiler JM, Olson J. Heparin detection by theactivated coagulation time: A comparison of thesensitivity of coagulation tests and heparin assays.J Cardiothorac Vasc Anesth 1997;11:24-28.

41. Gallus AS, Hirsh J, Tuttle RJ, Trebilcock R, O’BrienSE, Carroll JJ, Minden JH, Hudecki SM. Small sub-cutaneous doses of heparin in prevention of venousthrombosis. N Engl J Med 1973;288:545-551.

42. Ruff RL, Dougherty JH. Complications of anticoag-ulation followed by anticoagulation. Stroke 1981;12:879-881.

43. Rao TL , El-Etr AA. Anticoagulation following place-ment of epidural and subarachnoid catheters: Anevaluation of neurologic sequelae. Anesthesiology1981;55:618-620.

44. Wildsmith JA , McClure JH. Anticoagulant drugs


and central nerve blockade [editorial]. Anaesthesia1991;46:613-614.

45. Wulf H. Epidural anaesthesia and spinal haema-toma. Can J Anaesth 1996;43:1260-1271.

46. Chaney MA. Intrathecal and epidural anesthesiaand analgesia for cardiac surgery. Anesth Analg1997;84:1211-1221.

47. Goldstein S, Dean D, Kim SJ, Cocozell K, Grofsik J,Silver P, Cody RP. A survey of spinal and epiduraltechniques in adult cardiac surgery. J CardiothoracVasc Anesth 2001;15:158-168.

48. Hammer GB, Ngo K, Macario A. A retrospectiveexamination of regional plus general anesthesia inchildren undergoing open heart surgery. AnesthAnalg 2000;90:1020-1024.

49. Peterson KL, DeCampli WM, Pike NA, Robbins RC,Reitz BA. A report of two hundred twenty cases ofregional anesthesia in pediatric cardiac surgery.Anesth Analg 2000;90:1014-1019.

50. Sanchez, R, Nygard, E. Epidural anesthesia in car-diac surgery: Is there an increased risk? J Cardiotho-rac Vasc Anesth 1998;12:170-173.

51. Ralley FE. Neuraxial anesthesia is contraindicated inpatients undergoing heparinization for surgery. Pro:neuraxial anesthesia should not be used in patientsundergoing heparinization for surgery. J Cardiotho-rac Vasc Anesth 1996;10:957-960.

52. Turnbull KW. Neuraxial anesthesia is contraindi-cated in patients undergoing heparinization for sur-gery. Con: neuraxial block is useful in patients un-dergoing heparinization for surgery. J CardiothoracVasc Anesth 1996;10:961-962.

53. Anderson MB, Kwong KF, Furst AJ, Salerno TA.Thoracic epidural anesthesia for coronary bypass vialeft anterior thoracotomy in the conscious patient.Eur J Cardiothorac Surg 2001;20:415-417.

54. Vanek T, Straka Z, Brucek P, Widimsky P. Thoracicepidural anesthesia for off-pump coronary arterybypass without intubation. Eur J Cardiothorac Surg2001;20:858-860.

55. Priestley MC, Cope L, Halliwell R, Gibson P, ChardRB, Skinner M, Klineberg PL. Thoracic epidural an-esthesia for cardiac surgery: The effects on trachealintubation time and length of hospital stay. AnesthAnalg 2002;94:275-282.

56. Ho AM, Chung DC, Joynt GM. Neuraxial blockadeand hematoma in cardiac surgery: Estimating therisk of a rare adverse event that has not (yet) oc-curred. Chest 2000;117:551-555.

57. Castellano JM, Durbin CG Jr. Epidural analgesiaand cardiac surgery: Worth the risk? Chest 2000;117:305-307.

58. Collins R, Scrimgeour A, Yusuf S, Peto R. Reductionin fatal pulmonary embolism and venous thrombo-sis by perioperative administration of subcutaneousheparin. Overview of results of randomized trials ingeneral, orthopedic, and urologic surgery. N EnglJ Med 1988;318:1162-1173.

59. Wille-Jorgensen P, Jorgensen LN, Rasmussen LS.Lumbar regional anaesthesia and prophylactic anti-

coagulant therapy. Is the combination safe? Anaes-thesia 1991;46:623-627.

60. Millar FA, Mackenzie A, Hutchison G, BannisterJ. Hemostasis-altering drugs and central neuralblock—A survey of anesthetic practice in Scotlandand the United Kingdom. Reg Anesth Pain Med 1996;21:529-533.

61. Rodgers A, Sage D, Futter M, Macmahon S. Atti-tudes and practices of New Zealand anaesthetistswith regard to epidural and subarachnoid anaesthe-sia. Anaesth Intensive Care 1996;24:79-86.

62. Greaves JD. Serious spinal cord injury due tohaematomyelia caused by spinal anaesthesia in apatient treated with low-dose heparin. Anaesthesia1997;52:150-154.

63. Sandhu, H, Morley-Forster, P, Spadafora, S. Epi-dural hematoma following epidural analgesia in apatient receiving unfractionated heparin for throm-boprophylaxis. Reg Anesth Pain Med 2000;25:72-75.

64. Cosmi B, Hirsh J. Low molecular weight heparins.Curr Opin Cardiol 1994;9:612-618.

65. Levine MN, Planes A, Hirsh J. The relationship be-tween anti-factor Xa level and clinical outcome inpatients receiving enoxaparin low molecular weightheparin to prevent deep vein thrombosis after hipreplacement. Thromb Haemost 1989;62:940-944.

66. Kessler CM, Esparraguera IM, Jacobs HM, Druy E,Fortune WP, Holloway DS, Giordano J, DavidsonBL. Monitoring the anticoagulant effects of a lowmolecular weight heparin preparation. Correlationof assays in orthopedic surgery patients receivingardeparin sodium for prophylaxis of deep venousthrombosis. Am J Clin Pathol 1995;103:642-648.

67. Holst J, Lindblad B, Bergqvist D, Garre K, Nielsen H,Hedner U, Ostergaard PB. Protamine neutralizationof intravenous and subcutaneous low-molecular-weight heparin (tinzaparin, Logiparin™). An exper-imental investigation in healthy volunteers. BloodCoagul Fibrinolysis 1994;5:795-803.

68. Lojewski B, Bacher P, Iqbal O, Walenga JM, Hop-pensteadt D, Leya F, Fareed J. Evaluation of hemo-static and fibrinolytic alterations associated withdaily administration of low-molecular-weight hep-arin for a 12-week period. Semin Thromb Hemost1995;21:228-239.

69. Tryba M. European practice guidelines: Thrombo-embolism prophylaxis and regional anesthesia. RegAnesth Pain Med 1998;23:178-182.

70. Bergqvist D, Lindblad B, Matzsch T. Low molecularweight heparin for thromboprophylaxis and epi-dural/spinal anaesthesia—Is there a risk? Acta An-aesthesiol Scand 1992;36:605-609.

71. Bergqvist D, Lindblad B, Matzsch T. Risk of combin-ing low molecular weight heparin for thrombopro-phylaxis and epidural or spinal anesthesia. SeminThromb Hemost 1993;19:147-151.

72. Tryba M und die Teilnehmer des Workshops uberhamostaseologische Probleme bei regionalanaes-thesien. Hamostaseologische Voraussetzungen zurDurchfuhrung von Regionalanaesthesien. Reg An-aesth 1989;12:127-131.


73. Ang ET, Khon S, Delefosse D, Galet C, Goldfarb G,Jolis P. [Incidence of local complications after spinalanesthesia in patients treated with low molecularweight heparin.] [French] Ann Fr Anesth Reanim1989;8:R163 (suppl).

74. Planes A, Vochelle N, Fagola M, Feret J, Bellaud M.Prevention of deep vein thrombosis after total hipreplacement. The effect of low-molecular-weightheparin with spinal and general anaesthesia. J BoneJoint Surg Br 1991;73:418-422.

75. Tryba M, Wedel DJ. Central neuraxial block andlow molecular weight heparin (enoxaparin): Les-sons learned from two different dosage regimes intwo continents. Acta Anaesthesiol Scand 1997;41:100-104.

76. Llau JV, de Andres J, Gomar C, Gomez A, Hidalgo F,Sahagun J, Torres LM. [Drugs that alter hemostasisand regional anesthetic techniques: Safety guide-lines. Consensus conference.] [Spanish] Rev EspAnestesiol Reanim 2001;48:270-278.

77. Gogarten W, Van Aken H, Wulf H, Klose R, Van-dermeulen E, Harenberg J. Regional anesthesia andthromboembolism prophylaxis/anticoagulation. An-aesthesiol Intensivmed 1997;12:623-628.

78. Horlocker TT, Heit JA. Low molecular weight hep-arin: Biochemistry, pharmacology, perioperativeprophylaxis regimens and guidelines for regionalanesthetic management. Anesth Analg 1997;85:874-885.

79. Schroeder DR. Statistics: Detecting a rare adversedrug reaction using spontaneous reports. Reg AnesthPain Med 1998;23:183-189.

80. Douketis JD, Kinnon K, Crowther MA. Anticoagu-lant effect at the time of epidural catheter removalin patients receiving twice-daily or once-daily low-molecular-weight heparin and continuous epiduralanalgesia after orthopedic surgery. Thromb Haemost2002;88:37-40.

81. Rodriguez-Arias CA, Martinez R, Mac-Veigh I, Ta-boada M. Epidural lumbar hematoma, epiduralanesthesia and low-molecular-weight heparins.[Spanish] Neurochirurgia 2001;12:509-512.

82. Litz RJ, Hubler M, Koch T, Albrecht M. Spinal-epidural hematoma following epidural anesthesia inthe presence of antiplatelet and heparin therapy.Anesthesiology 2001;95:1031-1033.

83. Hull RD, Pineo GF, Francis C, Bergqvist D, FelleniusC, Soderberg K, Holmqvist A, Mant M, Dear R,Baylis B, Mah A, Brant R. Low-molecular-weightheparin prophylaxis using dalteparin in close prox-imity to surgery vs warfarin in hip arthroplasty pa-tients: A double-blind, randomized comparison. TheNorth American Fragmin Trial Investigators. ArchIntern Med 2000;160:2199-207.

84. Heit JA. Perioperative management of the chroni-cally anticoagulated patient. [Review] J ThrombosisThrombolysis 2001;12:81-87.

85. McEvoy MD, Bailey M, Taylor D, Del Schutte H.Noncompliance in the inpatient administration ofenoxaparin in conjunction with epidural or spinalanesthesia. J Arthroplasty 2000;15:604-607.

86. Majerus PW, Broze GJ, Miletich JP, Tollefsen DM:Anticoagulant, thrombolytic, and antiplatelet drugs.In: Hardman JG, Limbird LE, Molinoff PB, RuddonRW, Gilman AJ (eds): Pharmacologic Basis of Thera-peutics, ed 9. New York, NY: Macmillan; 1990:1341-1359.

87. Xi M, Beguin S, Hemker HC. The relative impor-tance of the Factors II, VII, IX, and X for the pro-thrombinase activity in plasma of orally anticoagu-lated patients. Thromb Haemost 1989;62:788-791.

88. Shields RC, McBane RD, Kuiper JD, Li H, Heit JA.Efficacy and safety of intravenous phytonadione(vitamin K1) in patients on long-term oral antico-agulant therapy. Mayo Clin Proc 2001;76:260-266.

89. Chan TYK. Adverse interactions between warfarinand nonsteroidal antiinflammatory drugs: Mecha-nisms, clinical significance, and avoidance. AnnPharmacother 1995;29:1274-1283.

90. Wille-Jorgensen P, Jorgensen LN, Rasmussen LS.Lumbar regional anesthesia and prophylactic anti-coagulant therapy: Is the combination safe. Anaes-thesia 1991;46:623-627.

91. Hirsh J: Oral anticoagulant drugs. N Engl J Med1991;324:1865-1875.

92. Harder S, Thurmann P. Clinically important druginteractions with anticoagulants: An update. ClinPharmacokinet 1996;30:416-444.

93. Warfarin labeling. Taro Pharmaceutical IndustriesLtd., Haifa Bay, Israel 26110. Last revised 1999.

94. Janis KM. Epidural hematoma following postoper-ative epidural analgesia: A case report. Anesth Analg1972;51:689-692.

95. Helperin SW, Cohen DD. Hematoma following epi-dural anesthesia: Report of a case. Anesthesiology1971;35:641-644.

96. Odoom JA, Sih IL. Epidural analgesia and anticoag-ulant therapy: Experience with one thousand casesof continuous epidurals. Anaesthesia 1983;38:254-259.

97. Horlocker TT, Wedel DJ, Schlicting JL. Postopera-tive epidural analgesia and oral anticoagulant ther-apy. Anesth Analg 1994;79:89-93.

98. Wu CL, Perkins FM. Oral anticoagulant prophylaxisand epidural catheter removal. Reg Anesth Pain Med1996;21:517-524.

99. Woolson ST, Robinson RK, Khan NQ, Rogers BS,Maloney WJ. Deep venous thrombosis prophylaxisfor knee replacement: Warfarin and pneumaticcompression. Am J Orthop 1998;27:299-304.

100. Badenhorst CH. Epidural hematoma after epiduralpain control and concomitant postoperative antico-agulation. Reg Anesth Pain Med 1996;21:272-273.

101. Rodgers RPC, Levin J. A critical reappraisal of thebleeding time. Semin Thromb Hemost 1990;16:1-20.

102. Cronberg S, Wallmark E, Soderberg I. Effect onplatelet aggregation of oral administration of 10non-steroidal analgesics to humans. Scand J Haema-tol 1984;33:155-159.

103. Leese PT, Hubbard RC, Karim A, Isakson PC, Yu SS,Geis GS. Effects of celecoxib, a novel cyclooxygen-ase-2 inhibitor, on platelet function in healthy


adults: A randomized, controlled trial. J Clin Phar-macol 2000;40:124-132.

104. Taniuchi M, Kurz HI, Lasala JM. Randomized com-parison of ticlopidine and clopidogrel after intra-coronary stent implantation in a broad patient pop-ulation. Circulation 2001;104:539-543.

105. Gerancher JC, Waterer R, Middleton J. Transientparaparesis after postdural puncture spinal hema-toma in a patient receiving ketorolac. Anesthesiology1997;86:490-494.

106. Benzon HT, Wong HY, Siddiqui T, Ondra S. Cautionin performing epidural injections in patients on sev-eral antiplatelet drugs. Anesthesiology 1999;91:1558-1559.

107. Gilbert A, Owens BD, Mulroy MF. Epidural hema-toma after outpatient epidural anesthesia. AnesthAnalg 2002;94:77-78.

108. Kawaguchi S, Tokutomi S. A case of epidural he-matoma associated with epidural catheterizationwhich occurred on 12th day after the last medica-tion of ticlopidine hydrochloride. [Japanese] Masui2002;51:526-528.

109. CLASP (Collaborative Low-Dose Aspirin Study inPregnancy) Collaborative Group. CLASP: A ran-domized trial of low-dose aspirin for the preventionand treatment of pre-eclampsia among 9364 preg-nant women. Lancet 1994;343:619-629.

110. Horlocker TT, Bajwa ZH, Ashraft Z, Khan S, WilsonJL, Sami N, Peeters-Asdourian C, Powers CA,Schroeder DR, Decker PA, Warfield CA. Risk assess-ment of hemorrhagic complications associated withnonsteroidal antiiflammatory medications in ambu-latory pain clinic patients undergoing epidural ste-roid injection. Anesth Analg 2002;95:1691-1697.

111. Coukell AJ. Clopidogrel. Drugs 1997;54:745-750.112. Dyke CM. Safety of glycoprotein IIb-IIIa inhibitors:

A heart surgeon’s perspective. [Review] Am Heart J1999;138:307-316.

113. Shlansky-Goldberg R. Platelet aggregation inhibi-tors for use in peripheral vascular interventions:What can we learn from the experience in the cor-onary arteries? [Review] J Vasc Interv Radiol 2002;13:229-246.

114. Mayumi T, Dohi S. Spinal subarachnoid hematomaafter lumbar puncture in a patient receiving anti-platelet therapy. Anesth Analg 1983:62:777-779.

115. Hensrud DD, Engle DD, Scheitel SM. Underreport-ing the use of dietary supplements and nonprescrip-tion medications among patients undergoing a pe-riodic health examination. Mayo Clin Proc 1999;74:443-447.

116. Elder NC, Gillcrist A, Minz R. Use of alternativehealth care by family practice patients. Arch FamMed 1997;6:181-184.

117. Eisenberg DM. Advising patients who seek alterna-tive medical therapies. Ann Intern Med 1997;127:61-69.

118. Stevinson C, Pittler MH, Ernst E. Garlic for treatinghypercholesterolemia: A meta-analysis of random-ized clinical trials. Ann Intern Med 2000;133:420-429.

119. [No author]. Garlic. In: Blumenthal M, Goldberg A,Brinckmann J, eds. Herbal Medicine–Expanded Com-mission E Monographs, ed 1. Newton, MA: IntegrativeMedical Communications; 2000:139-148.

120. Srivastava KC. Evidence for the mechanism bywhich garlic inhibits platelet aggregation. Prostaglan-dins Leukot Med 1986;22:313-321.

121. Apitz-Castro R, Escalante J, Vargas R, Jain MK.Ajoene, the antiplatelet principle of garlic, synergis-tically potentiates the antiaggregatory action ofprostacyclin, forskolin, indomethacin and dypiri-damole on human platelets. Thromb Res 1986;42:303-311.

122. Rose KD, Croissant PD, Parliament CF, Levin MB.Spontaneous spinal epidural hematoma with asso-ciated platelet dysfunction from excessive garlic in-gestion: A case report. Neurosurgery 1990:26:880-882.

123. Le Bars PL, Katz MM, Berman N, Itil TM, FreedmanAM, Schatzberg AF. A placebo-controlled, double-blind, randomized trial of an extract of Ginkgo bi-loba for dementia. North American Egb StudyGroup. JAMA 1997;278:1327-1332.

124. Oken BS, Storzbach DM, Kaye JA. The efficacy ofGinkgo biloba on cognitive function in Alzheimerdisease. Arch Neurol 1998;55:1409-1415.

125. [No author]. Ginkgo biloba leaf extract. In: Blumen-thal M, Goldberg A, Brinckmann J, eds. Herbal Medi-cine–Expanded Commission E Monographs, ed 1. New-ton, MA: Integrative Medical Communications;2000:160-169.

126. Chung KF, Dent G, McCusker M, Guinot P, PageCP, Barnes PJ. Effect of a ginkgolide mixture (BN52063) in antagonizing skin and platelet responsesto platelet activating factor in man. Lancet 1987;1:248-251.

127. Rowin J, Lewis SL. Spontaneous bilateral subduralhematomas associated with chronic ginkgo bilobaingestion. Neurology 1996;46:1775-1776.

128. Vale S. Subarachnoid haemorrhage associated withGinkgo biloba. Lancet 1998;352:36.

129. Gilbert GJ. Ginkgo biloba. Neurology 1997;48:1137.130. Matthews MK Jr. Association of Ginkgo biloba with

intracerebral hemorrhage. Neurology 1998;50:1933-1934.

131. Brekham II, Dardymov IV. New substances of plantorigin which increase nonspecific resistance. AnnuRev Pharmacol 1969;9:419-430.

132. [No author]. Ginseng root. In: Blumenthal M, Gold-berg A, Brinckmann J, eds. Herbal Medicine–ExpandedCommission E Monographs, ed 1. Newton, MA: Integra-tive Medical Communications; 2000:170-177.

133. Attele AS, Wu JA, Yuan CS. Ginseng pharmacology:Multiple constituents and multiple actions. BiochemPharmacol 1999;58:1685-1693.

134. Kimura Y, Okuda H, Arichi S. Effects of variousginseng saponins on 5-hydroxytryptamine releaseand aggregation in human platelets. J Pharm Phar-macol 1988;40:838-843.

135. Kuo SC, Teng CM, Lee JC, Ko FN, Chen SC, Wu TS.


Antiplatelet components in Panax ginseng. PlantaMed 1990;56:164-167.

136. Park HJ, Lee JH, Song YB, Park KH. Effects of di-etary supplementation of lipophilic fraction fromPanax ginseng on cGMP and cAMP in rat plateletsand on blood coagulation. Biol Pharm Bull 1996;19:1434-1439.

137. Janetzky K, Morreale AP. Probable interaction be-tween warfarin and ginseng. Am J Health Syst Pharm1997;54:692-693.

138. Carswell CI, Plosker GL. Bivalirudin: A review of itspotential place in the management of acute coro-nary syndromes. [Review] Drugs 2002;62:841-870.

139. Greinacher A, Lubenow N. Recombinant hirudin inclinical practice: Focus on lepirudin. [Review] Cir-culation 2001;103:1479-1484.

140. Eriksson BI, Wille-Jorgensen P, Kalebo P, Mouret P,Rosencher N, Bosch P, Baur M, Ekman S, Bach D,Lindbratt S, Close P. A comparison of recombinanthirudin with a low-molecular-weight heparin toprevent thromboembolic complications after totalhip replacement. N Engl J Med 1997;337:1329-1335.

141. Turpie AG, Gallus AS, Hoek JA. PentasaccharideInvestigators. A synthetic pentasaccharide for theprevention of deep-vein thrombosis after total hipreplacement. N Engl J Med 2001;344:619-625.

142. Landow L. A synthetic pentasaccharide for the pre-vention of deep-vein thrombosis. [Letter] N EnglJ Med 2001;345:291-292.

143. Turpie AG, Gallus AS, Hoek JA. Pentasaccharide

Investigators. A synthetic pentasaccharide for theprevention of deep-vein thrombosis after total hipreplacement. [Letter] N Engl J Med 2001;345:292.

144. Keegan MT, Horlocker TT. Epidural catheter re-moval before unanticipated anticoagulation: Thepharmacy fail-safe. [Letter] Anesthesiology 1999;91:328.

145. Fransson SG, Nylander E. Vascular injury followingcardiac catheterization, coronary angiography, andcoronary angioplasty. Eur Heart J 1994;15:232-235.

146. Johr M. A complication of continuous blockade ofthe femoral nerve. [German] Regional Anaesthesie1987;10:37-78.

147. Ben-David B, Stahl S. Axillary block complicated byhematoma and radial nerve injury. Reg Anesth PainMed 1999;24:264-266.

148. Vaisman J. Pelvic hematoma after an ilioinguinalnerve block for orchialgia. Anesth Analg 2001;92:1048-1049.

149. Ekatodramis G, Macaire P, Borgeat A. ProlongedHorner syndrome due to neck hematoma after con-tinuous interscalene block. Anesthesiology 2001;95:801-803.

150. Klein SM, D’Ercole F, Greengrass RA, Warner DS.Enoxaparin associated with psoas hematoma andlumbar plexopathy after lumbar plexus block. Anes-thesiology 1997;87:1576-1579.

151. Nielsen CH. Bleeding after intercostal nerve block ina patient anticoagulated with heparin. Anesthesiology1989;71:162-164.


Regional Anesthesia in the Anticoagulated Patient:...

Documents

Transcript of Regional Anesthesia in the Anticoagulated Patient:...