European Journal of Pain Supplements 3 (2009) 49–54

Contents lists available at ScienceDirect

European Journal of Pain Supplements

journal homepage: www.EuropeanJournalPain.com

Locoregional anesthesia and anticoagulation

Terese T. Horlocker *

Department of Anesthesiology, Mayo Clinic, 200 First Street Southwest, Rochester, MN 55905, United States

a r t i c l e i n f o

Article history:Received 17 June 2009Accepted 21 July 2009

Keywords:Neuraxial blockadeSpinal anesthesiaEpidural anesthesiaComplicationsParalysisHematomaAnticoagulation

1754-3207/$36.00 � 2009 European Federation of Intdoi:10.1016/j.eujps.2009.07.001

* Tel.: +1 507 284 9694; fax: +1 507 284 0120.E-mail address: horlocker.terese@mayo.edu

a b s t r a c t

Spinal hematoma is a rare and potentially catastrophic complication of spinal or epidural anesthesia. Riskfactors include traumatic needle/catheter placement, sustained anticoagulation in an indwelling neurax-ial catheter, and catheter removal during therapeutic levels of anticoagulation. Generally, a patient’scoagulation status should be optimized at the time of spinal or epidural needle/catheter placement,and the level of anticoagulation should be monitored during epidural catheterization. Signs of cord com-pression, such as severe back pain, progression of numbness or weakness, and bowel and bladder dys-function, warrant immediate radiographic evaluation. A delay in diagnosis and intervention of spinalhematoma may lead to irreversible cord ischemia.

� 2009 European Federation of International Association for the Study of Pain Chapters. Published byElsevier Ltd. All rights reserved.

1. Introduction

Spinal hematoma, defined as symptomatic bleeding within thespinal neuraxis, is a rare and potentially catastrophic complicationof spinal or epidural anesthesia. The actual incidence of neurologicdysfunction resulting from hemorrhagic complications associatedwith central neural blockade is unknown. In an extensive reviewof the literature, Tryba identified 13 cases of spinal hematoma fol-lowing 850,000 epidural anesthetics and 7 cases among 650,000spinal techniques (Tryba, 1993). Based on these observations thecalculated incidence is approximated to be less than 1 in 150,000epidural and less than 1 in 220,000 spinal anesthetics. Since theseestimates represent the upper limit of the 95% confidence interval,the actual frequency may be much less. Hemorrhage into the spinalcanal most commonly occurs in the epidural space, most likely be-cause of the prominent epidural venous plexus, although anes-thetic variables, such as needle size and catheter placement mayalso affect the site of clinically significant bleeding.

In a review of the literature between 1906 and 1994, Van-dermeulen et al. (1994) reported 61 cases of spinal hematomaassociated with epidural or spinal anesthesia. In 42 of the 61 pa-tients (68%), the spinal hematomas associated with central neuralblockade occurred in patients with evidence of hemostatic abnor-mality. Twenty-five of the patients had received intravenous orsubcutaneous (unfractionated or low molecular weight) heparin,while an additional five patients were presumably administeredheparin as they were undergoing a vascular surgical procedure.

ernational Association for the Stud

In addition, 12 patients had evidence of coagulopathy or thrombo-cytopenia or were treated with antiplatelet medications (aspirin,indomethacin, ticlopidine), oral anticoagulants (phenprocoumone),thrombolytics (urokinase), or dextran 70 immediately before orafter the spinal or epidural anesthetic. Needle and catheter place-ment was reported to be difficult in 15 (25%), or bloody in 15(25%) patients. Overall, in 53 of the 61 cases (87%), either a clottingabnormality or needle placement difficulty was present. A spinalanesthetic was performed in 15 patients. The remaining 46 pa-tients received an epidural anesthetic, including 32 patients withan indwelling catheter. In 15 of these 32 patients, the spinal hema-toma occurred immediately after the removal of the epidural cath-eter. Nine of these catheters were removed during therapeuticlevels of heparinization. Neurologic compromise presented as pro-gression of sensory or motor block (68% of patients) or bowel/blad-der dysfunction (8% of patients), not severe radicular back pain.Importantly, although only 38% of patients had partial or good neu-rologic recovery, spinal cord ischemia tended to be reversible inpatients who underwent laminectomy within 8 h of onset of neu-rologic dysfunction (Vandermeulen et al., 1994) (Table 1).

The need for prompt diagnosis and intervention in the event ofa spinal hematoma was also demonstrated in a recent review of theAmerican Society of Anesthesiologists (ASA) Closed Claims data-base, which noted that spinal cord injuries were the leading causeof claims in the 1990s (Cheney et al., 1999). Spinal hematomas ac-counted for nearly half of the spinal cord injuries. Risk factors forspinal hematoma included epidural anesthesia in the presence ofintravenous heparin during a vascular surgical or diagnostic proce-dure. Importantly, the presence of postoperative numbness orweakness was typically attributed to local anesthetic effect rather

y of Pain Chapters. Published by Elsevier Ltd. All rights reserved.

Table 1Neurologic outcome in patients with spinal hematoma following neuraxial blockade.a

Interval between onset of paraplegiaand surgery

GoodN = 15

PartialN = 11

PoorN = 29

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

Adapted from Vandermeulen et al. (1994). With permission.a Neurologic outcome was reported for 55 of 61 cases of spinal hematoma fol-

lowing neuraxial blockade.

50 T.T. Horlocker / European Journal of Pain Supplements 3 (2009) 49–54

than spinal cord ischemia, which delayed the diagnosis. Patientcare was rarely judged to have met standards (1 of 13 cases) andthe median payment was very high.

It is impossible to conclusively determine risk factors for thedevelopment of spinal hematoma in patients undergoing neuraxialblockade solely through review of the case series, which representonly patients with the complication and do not define those whounderwent uneventful neuraxial analgesia. However, large inclu-sive surveys that evaluate the frequencies of complications(including spinal hematoma), as well as identify subgroups of pa-tients with higher or lower risk, enhance risk stratification. Moenet al. (2004) investigated serious neurologic complications among1,260,000 spinal and 450,000 epidural blocks performed in Swedenover a 10-year period. Among the 33 spinal hematomas, 24 oc-curred in females; 25 were associated with an epidural technique.The methodology allowed for calculation of frequency of spinalhematoma among patient populations. For example, the risk asso-ciated with epidural analgesia in women undergoing childbirthwas significantly less (1 in 200,000) than that in elderly womenundergoing knee arthroplasty (1 in 3600, p < 0.0001). Likewise, wo-men undergoing hip fracture surgery under spinal anesthesia hadan increased risk of spinal hematoma (1 in 22,000) compared toall patients undergoing spinal anesthesia (1 in 480,000).

Overall, these series suggest that the risk of clinically significantbleeding varies with age (and associated abnormalities of thespinal cord or vertebral column), the presence of an underlyingcoagulopathy, difficulty during needle placement, and an indwell-ing neuraxial catheter during sustained anticoagulation (particu-larly with standard heparin or LMWH); perhaps in amultifactorial manner. They also consistently demonstrate theneed for prompt diagnosis and intervention.

2. Oral anticoagulants

Few data exist regarding the risk of spinal hematoma in patientswith indwelling epidural catheters who are anticoagulated withwarfarin. The optimal duration of an indwelling catheter and thetiming of its removal also remain controversial. To date, only threestudies have evaluated the risk of spinal hematoma in patients withindwelling spinal or epidural catheters who receive oral anticoagu-lants perioperatively. Odoom and Sih (1983) performed 1000 con-tinuous lumbar epidural anesthetics in vascular surgical patientswho were receiving oral anticoagulants preoperatively. The throm-botest (a test measuring factor IX activity) was decreased in all pa-tients prior to needle placement. Heparin was also administeredintraoperatively. Epidural catheters remained in place for 48 hpostoperatively. There were no neurologic complications. Whilethese results are reassuring, the obsolescence of the thrombotestas a measure of anticoagulation combined with the unknown coag-ulation status of the patients at the time of catheter removal limitthe usefulness of these results. Therefore, except in extraordinarycircumstances, spinal or epidural needle/catheter placement andremoval should not be performed in fully anticoagulated patients.

There were also no symptomatic spinal hematomas in 192 pa-tients receiving postoperative epidural analgesia in conjunctionwith low-dose warfarin after total knee arthroplasty. Patients re-ceived warfarin, starting on the postoperative day, to prolong theprothrombin time (PT) to 15.0–17.3 s (normal 10.9–12.8 s), corre-sponding to an international normalized ratio (INR) of 2.0–3.0. Epi-dural catheters were left indwelling 37 ± 15 h (range 13–96 h).Mean PT at the time of epidural catheter removal was 13.4 ± 2 s(range 10.6–25.8 s). This study documents the relative safety oflow-dose warfarin anticoagulation in patients with an indwellingepidural catheter. It also demonstrates the large variability in pa-tient response to warfarin. The authors recommended close moni-toring of coagulation status to avoid excessive prolongation of thePT during epidural catheterization (Horlocker et al., 1994).

Wu and Perkins (1996) retrospectively reviewed the medical re-cords of 459 patients who underwent orthopedic surgical proce-dures under spinal or epidural anesthesia, including 412 patientswho received postoperative epidural analgesia. Prothrombin timeat the time of epidural catheter removal was 14.1 ± 3.2 s (normal9.6–11.1 s), corresponding to an INR of 1.4. Patients who had war-farin thromboprophylaxis initiated preoperatively had signifi-cantly higher PTs at the time of catheter removal than patientswho had received postoperative warfarin only.

2.1. Regional anesthetic management of the patient on oralanticoagulants

Anesthetic management of patients anticoagulated periopera-tively with warfarin is dependent on dosage and timing of initia-tion of therapy. Since factor VII has a relatively short half-life,prolongation of the PT and INR may occur in 24–36 h after initia-tion of warfarin therapy. The PT will be prolonged (outside of nor-mal range) when factor VII activity is reduced to approximately55% of baseline. However, the therapeutic effect of warfarin antico-agulation is most dependent on reduction in factors II and X activ-ity. Since these factors have circulating half-lives of 36–48 and 72–96 h, respectively, thromboprophylaxis is not adequate for 3–5 days after starting warfarin therapy (Horlocker et al., 2003).

Many orthopedic surgeons administer the first dose of warfarinthe night before surgery. For these patients, the PT and INR shouldbe checked prior to neuraxial block if the first dose was given morethan 24 h earlier, or a second dose of oral anticoagulant has beenadministered. Patients receiving low-dose warfarin therapy duringepidural analgesia should have their PT and INR monitored on adaily basis, and checked before catheter removal, if initial dose ofwarfarin was more than 36 h beforehand. Reduced doses of warfa-rin should be given to patients who are likely to have an enhancedresponse to the drug. In general, is it recommended that indwellingneuraxial catheters be removed when the INR < 1.5 in order to as-sure adequate levels of all vitamin K-dependent factors. An INR > 3should prompt the physician to withhold or reduce the warfarindose in patients with indwelling neuraxial catheters. There is nodefinitive recommendation for removal of neuraxial catheters inpatients with therapeutic levels of anticoagulation during a neur-axial catheter infusion (Horlocker et al., 2003).

The PT and INR of patients on chronic oral anticoagulation willrequire 3–5 days to normalize after warfarin discontinuation. The-oretically, since the PT and INR reflect predominantly factor VIIactivity (and factor VII has only a 6–8 h half-life), there may bean interval during which the PT and INR approach normal values,yet factors II and X levels may not be adequate for normal hemos-tasis. Adequate levels of all vitamin K-dependent factors are typi-cally present when the INR is in the normal range. Therefore, it isrecommended that documentation of the patient’s normal coagu-lation status be achieved prior to implementation of neuraxialblock (Horlocker et al., 1994, 2003).

T.T. Horlocker / European Journal of Pain Supplements 3 (2009) 49–54 51

3. Intravenous and subcutaneous standard heparin

Complete systemic heparinization is typically reserved for themost high-risk patients, typically patients with an acute thrombo-embolism. However, intraoperative administration of a modestintravenous dose is occasionally performed during vascular ororthopedic procedures. In a study involving over 4000 patients,Rao and El-Etr (1981) demonstrated the safety of indwelling spinaland epidural catheters during systemic heparinization. However,the heparin activity was closely monitored, the indwelling cathe-ters were removed at a time when circulating heparin levels wererelatively low, and patients with a preexisting coagulation disorderwere excluded. A subsequent study in the neurologic literature byRuff and Dougherty (1981) reported spinal hematomas in 7 of 342patients (2%) who underwent a diagnostic lumbar puncture andsubsequent heparinization. Traumatic needle placement, initiationof anticoagulation within 1 h of lumbar puncture or concomitantaspirin therapy were identified as risk factors in the developmentof spinal hematoma in anticoagulated patients. Overall, large pub-lished series and extensive clinical experience suggests the use ofregional techniques during systemic heparinization does not ap-pear to represent a significant risk. However, the recent reportsof paralysis relating to spinal hematoma in the ASA Closed Claimsdatabase suggests that these events may not be as rare as sus-pected and that extreme vigilance is necessary to diagnose andintervene as early as possible, should spinal hematoma be sus-pected (Cheney et al., 1999).

The use of epidural and spinal anesthesia and analgesia in thepresence of high dose intraoperative systemic heparin, specificallyin cardiac surgery has gained recent popularity. In a recent surveyof the membership of the Society of Cardiovascular Anesthesiolo-gists, Goldstein et al. (2001) surveyed 3974 cardiac anesthesiolo-gists, and found 7% of their responders used spinal or epiduraltechniques for cardiac surgery. Interestingly, the majority of anes-thesiologists would proceed if frank blood was noted in the spinalor epidural needle. To date there are no case reports of spinalhematomas associated with this technique published or withinthe Closed Claims Project (Cheney et al., 1999). Ho et al. calculatedthe risk of hematoma among these patients. In a complex mathe-matical analysis of the probability of predicting a rare event thathas not occurred yet, they estimate the probability of an epiduralhematoma (based on the totals of 4583 epidural and 10,840 spinalanesthetics reported without complications) to be in the neighbor-hood of 1:1528 for epidural injection, and 1:3610 for spinal tech-nique (Ho et al., 2000).

Low-dose subcutaneous standard (unfractionated) heparin isadministered for thromboprophylaxis in patients undergoing ma-jor thoracoabdominal surgery and in patients at increased risk ofhemorrhage with oral anticoagulant or LMWH therapy. As previ-ously mentioned, subcutaneous heparin does not provide ade-quate prophylaxis following major orthopedic surgery, and isseldom utilized in this patient population. A review of the litera-ture by Schwander and Bachmann (1991) noted no spinal hema-tomas in over 5000 patients who received subcutaneous heparinin combination with spinal or epidural anesthesia. There are onlythree cases of spinal hematoma associated with neuraxial block-ade in the presence of low-dose heparin, two of which involveda continuous epidural anesthetic technique (Vandermeulenet al., 1994).

3.1. Regional anesthetic management of the patient receiving standardheparin

The safety of neuraxial techniques in combination with intraop-erative heparinization is well documented, providing no other

coagulopathy is present. The concurrent use of medications that af-fect other components of the clotting mechanisms may increasethe risk of bleeding complications for patients receiving standardheparin. These medications include antiplatelet medications,LMWH, and oral anticoagulants (Horlocker et al., 2003).

Intravenous heparin administration should be delayed for 1 hafter needle placement. Indwelling catheters should be removed1 h before a subsequent heparin administration or 2–4 h after thelast heparin dose. Evaluation of the coagulation status may beappropriate prior to catheter removal in patients who have demon-strated enhanced response or are on higher doses of heparin.Although the occurrence of a bloody or difficult needle placementmay increase risk, there are no data to support mandatory cancel-lation of a case should this occur. If the decision is made to pro-ceed, full discussion with the surgeon and careful postoperativemonitoring are warranted (Horlocker et al., 2003).

Prolonged therapeutic anticoagulation appears to increase riskof spinal hematoma formation, especially if combined with otheranticoagulants or thrombolytics. Therefore, neuraxial blocksshould be avoided in this clinical setting. If systematic anticoagula-tion therapy is begun with an epidural catheter in place, it is rec-ommended to delay catheter removal for 2–4 h following heparindiscontinuation and after evaluation of coagulation status (Hor-locker et al., 2003).

There is no contradiction to use of neuraxial techniques duringsubcutaneous standard heparin 610,000 U/day. However, higherdoses are associated with increased medical and surgical bleedingand may also increase the risk of spinal hematoma. Since the risk ofspinal hematoma is unknown in patients receiving 5000 U TID or7500 U BID dosing, the risks and benefits should be assessed onan individual basis. The risk of neuraxial bleeding may be reducedby delay of the heparin injection until after the block, and may beincreased in debilitated patients or after prolonged therapy. Aplatelet count is indicated for patients receiving subcutaneous hep-arin for greater than 5 days (Horlocker et al., 2003).

4. Low molecular weight heparin

Enoxaparin, the first LMWH to be approved by the Food andDrug Administration (FDA) in the United States, was distributedfor general use in May 1993. Within 1 year, two cases of spinalhematoma had been voluntarily reported through the MedWatchsystem. Despite repeated efforts at relabeling and education, casesof spinal hematoma continued to occur. A total of 30 cases ofspinal hematoma in patients undergoing spinal or epidural anes-thesia while receiving LMWH perioperatively were reported be-tween May 1993 and November 1997. An FDA Health Advisorywas issued in December 1997. In addition, the manufacturers ofall LMWH and heparinoids were requested to place a black‘‘boxed warning”.

At the time of the Consensus Conference on Neuraxial Anesthe-sia and Anticoagulation on April 1998, there were 45 cases ofspinal hematoma associated with LMWH, 40 involved a neuraxialanesthetic. Severe radicular back pain was not the presentingsymptom; most patients complained of new onset numbness,weakness, or bowel and bladder dysfunction. Median time intervalbetween initiation of LMWH therapy and neurologic dysfunctionwas 3 days, while median time to onset of symptoms and laminec-tomy was over 24 h. Less than one third of the patients reportedfair or good neurologic recovery (Horlocker and Wedel, 1998).

The risk of spinal hematoma, based on LMWH sales, prevalenceof neuraxial techniques and reported cases, was estimated to beapproximately 1 in 3000 continuous epidural anesthetics com-pared to 1 in 40,000 spinal anesthetics (Schroeder, 1998). However,this is most likely an underestimation – in addition to the spinal

52 T.T. Horlocker / European Journal of Pain Supplements 3 (2009) 49–54

hematomas that had been reported at the time of the First Consen-sus Conference, there were approximately 20 more that hadoccurred, but were not yet reported to the MedWatch system. Intotal, nearly 60 spinal hematomas were tallied by the FDA between1993 and 1998.

There have been only 13 cases of spinal hematoma followingneuraxial block since 1998 reported through the MedWatch sys-tem or published as case reports. In addition to LMWH, five pa-tients received ketorolac, one patient received ibuprofen, and onepatient received intravenous unfractionated heparin during a vas-cular procedure. The regional technique was a spinal anesthetic inthree cases. The remaining 10 patients underwent epidural anes-thesia in combination with LWMH therapy. Thus, the characteris-tics of the reported cases support the previous recommendationsof epidural catheter removal prior to the initiation of LMWHthromboprophylaxis and avoidance of concomitant antiplatelet/anticoagulant medications (Table 2). Although the number of casesvoluntarily reported has markedly declined, this may be a result ofdecreased reporting, improved management, or simple avoidanceof all neuraxial techniques in patients receiving LMWH. Continuedmonitoring is necessary.

The indications and labeled uses for LMWH continue to evolve.Indications for thromboprophylaxis as well as treatment of DVT/PEor MI have been introduced since the first Consensus Conference.These new applications and corresponding regional anestheticmanagement warrant discussion (Geerts et al., 2004). Several off-label applications of LMWH are of special interest to the anesthe-siologist. LMWH has been demonstrated to be efficacious as a‘‘bridge therapy” for patients chronically anticoagulated with war-farin, including parturients, patients with prosthetic cardiac valves,a history of atrial fibrillation, or preexisting hypercoagulable condi-tion. The doses of LMWH are those associated with DVT treatment,not prophylaxis, and are much higher. Needle placement shouldoccur a minimum of 24 h following this level of LMWHanticoagulation.

4.1. Regional anesthetic management of the patient receiving lowmolecular weight heparin

Perioperative management of patients receiving LMWH re-quires coordination and communication. It is also important tonote that even when protocols for dosing of LMWH and cathetermanagement exist, they may not be closely followed. McEvoyet al. (2000) reported a 52% noncompliance rate in the administra-tion of LMWH in association with epidural analgesia. Clinicians areurged to develop protocols that ‘‘fit” within the normal practicestandards at their institution, rather than deviate from the routine.Antiplatelet or oral anticoagulant medications administered incombination with LMWH increase the risk of spinal hematoma.

Table 2Patient, anesthetic, and low molecular weight heparin (LMWH) dosing variablesassociated 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

From Horlocker and Wedel (1998). With permission.

The presence of blood during needle and catheter placement doesnot necessitate postponement of surgery (Horlocker et al., 2003).

4.1.1. Preoperative LMWHPatients on preoperative LMWH can be assumed to have altered

coagulation. A single-injection spinal anesthetic may be the safestneuraxial technique in patients receiving preoperative LMWH forthromboprophylaxis. In these patients needle placement shouldoccur at least 10–12 h after the LMWH dose. Patients receivingtreatment doses of LMWH will require delays of at least 24 h.Neuraxial techniques should be avoided in patients administereda dose of LMWH 2 h preoperatively (general surgery patients), be-cause needle placement would occur during peak anticoagulantactivity (Horlocker et al., 2003).

4.1.2. Postoperative LMWHPatients with postoperative initiation of LMWH thrombopro-

phylaxis may safely undergo single-injection and continuous cath-eter techniques. Management is based on total daily dose, timingof the first postoperative dose and dosing schedule. The first doseof LMWH should be administered no earlier than 24 h postopera-tively, regardless of anesthetic technique, and only in the presenceof adequate hemostasis. LMWH may be administered in the pres-ence of an indwelling catheter only if LMWH is administered oncedaily and no additional hemostasis-altering medications areadministered (e.g. ketorolac). However, ideally indwelling cathe-ters should be removed prior to initiation of LMWH thrombopro-phylaxis (Horlocker et al., 2003).

5. Antiplatelet medications

Antiplatelet medications are seldom used as primary agents ofthromboprophylaxis. However, many orthopedic patients reportchronic use of one or more antiplatelet drugs (Horlocker et al.,1995). Although Vandermeulen et al. (1994) implicated anti-platelet therapy in 3 of the 61 cases of spinal hematoma occur-ring after spinal or epidural anesthesia, several large studieshave demonstrated the relative safety of neuraxial blockade inboth obstetric, surgical and pain clinic patients receiving thesemedications (CLASP Collaborative Group, 1994; Horlocker et al.,1995, 2002). In a prospective study involving 1000 patients, Hor-locker et al. (1995) reported that preoperative antiplatelet ther-apy did not increase the incidence of blood present at the timeof needle/catheter placement or removal, suggesting that traumaincurred during needle or catheter placement is neither increasednor sustained by these medications. The clinician should beaware of the possible increased risk of spinal hematoma in pa-tients receiving antiplatelet medications who undergo subse-quent heparinization (Ruff and Dougherty, 1981). Ticlopidineand clopidogrel are also platelet aggregation inhibitors. Theseagents interfere with platelet–fibrinogen binding and subsequentplatelet–platelet interactions. The effect is irreversible for the lifeof the platelet. Ticlopidine and clopidogrel have no effect onplatelet cyclooxygenase, acting independently of aspirin. How-ever, these medications have not been tested in combination.Platelet dysfunction is present for 5–7 days after discontinuationof clopidogrel and 10–14 days with ticlopidine. Platelet glycopro-tein IIb/IIIa receptor antagonists, including abciximab (Reopro�),eptifibatide (Integrilin�) and tirofiban (Aggrastat�), inhibit plate-let aggregation by interfering with platelet–fibrinogen bindingand subsequent platelet–platelet interactions. Time to normalplatelet aggregation following discontinuation of therapy rangesfrom 8 h (eptifibatide, tirofiban) to 48 h (abciximab). Increasedperioperative bleeding in patients undergoing cardiac and vascu-lar surgery after receiving ticlopidine, clopidogrel and glycopro-

T.T. Horlocker / European Journal of Pain Supplements 3 (2009) 49–54 53

tein IIb/IIIa antagonists (Kovesi and Royston, 2002) warrants con-cern regarding the risk of anesthesia-related hemorrhagiccomplications.

5.1. Regional anesthetic management of the patient receivingantiplatelet medications

Antiplatelet drugs, by themselves, appear to represent no addedsignificant risk for the development of spinal hematoma in patientshaving epidural or spinal anesthesia. However, the concurrent useof medications that affect other components of the clotting mech-anisms, such as oral anticoagulants, standard heparin, and LMWH,may increase the risk of bleeding complications for patients receiv-ing antiplatelet agents. Assessment of platelet function prior toperformance of neuraxial block is not recommended. However,careful preoperative assessment of the patient to identify altera-tions of health that might contribute to bleeding is crucial (Hor-locker et al., 2003).

The increase in perioperative bleeding in patients undergoingcardiac and vascular surgery after receiving ticlopidine, clopidogreland platelet GP IIb/IIIa antagonists warrants concern regarding therisk of spinal hematoma. The recommended time interval betweendiscontinuation of thienopyridine therapy and neuraxial blockadeis 14 days for ticlopidine and 7 days for clopidogrel. For the plateletGP IIb/IIIa inhibitors, the duration ranges from 8 h for eptifibatideand tirofiban, to 48 h following abciximab administration (Hor-locker et al., 2003).

6. Herbal medications

There is a widespread use of herbal medications in surgical pa-tients. Morbidity and mortality associated with herbal use may bemore likely in the perioperative period because of the polyphar-macy and physiological alterations that occur. Such complicationsinclude bleeding from garlic, ginkgo and ginseng, and potentialinteraction between ginseng and warfarin. Because the currentregulatory mechanism for commercial herbal preparations sold inthe United States does not adequately protect against unpredict-able or undesirable pharmacological effects, it is especially impor-tant for anesthesiologists to be familiar with related literature onherbal medications when caring for patients in the perioperativeperiod (Rose et al., 1990).

Table 3Recommendations for management of patients receiving neuraxial blockade and anticoag

Warfarin Discontinue chronic warfarin therapy 4–5 days before spinaprocedure to ensure adequate levels of all vitamin K-dependewith INR < 1.5

Antiplateletmedications

No contraindications with aspirin or other NSAIDs. Thienopy14 days, respectively, prior to procedure. GP IIb/IIIa inhibitor(8 h for tirofiban and eptifibatide, 24–48 h for abciximab)

Thrombolytics/fibrinolytics

There are no available data to suggest a safe interval betweefibrinogen level and observe for signs of neural compression

LMWH Delay procedure at least 12 h from the last dose of thromboelapse prior to procedure. LMWH should not be administeremaintained with caution and only with once daily dosing ofincluding ketorolac

Unfractionated SQheparin

There are no contraindications to neuraxial procedure if totaunknown; assess on an individual basis.

Unfractionated IVheparin

Delay needle/catheter placement 2–4 h after last dose, docuheparinization with an indwelling neuraxial catheter associa

NSAIDs, nonsteroidal antiinflammatory drugs; GP IIb/IIIa, platelet glycoprotein receptor Iheparin; aPTT, activated partial thromboplastin time.Adapted from Horlocker et al. (2003).

6.1. Regional anesthetic management of the patient receiving herbaltherapy

Herbal drugs, by themselves, appear to represent no added sig-nificant risk for the development of spinal hematoma in patientshaving epidural or spinal anesthesia. This is an important observa-tion since it is likely that a significant number of our surgical pa-tients utilize alternative medications preoperatively and perhapsduring their postoperative course. There is no wholly accepted testto assess adequacy of hemostasis in the patient reporting preoper-ative herbal medications. Careful preoperative assessment of thepatient to identify alterations of health that might contribute tobleeding is crucial. Data on the combination of herbal therapy withother forms of anticoagulation are lacking. However, the concur-rent use of other medications affecting clotting mechanisms mayincrease the risk of bleeding complications in these patients (Hor-locker et al., 2003).

7. Plexus and peripheral blockade in the anticoagulated patient

Although spinal hematoma is the most significant hemorrhagiccomplication of regional anesthesia due to the catastrophic natureof bleeding into a fixed and noncompressible space, the associatedrisk following plexus and peripheral techniques remains unde-fined. There are few investigations which examine the frequencyand severity of hemorrhagic complications following plexus orperipheral blockade in anticoagulated patients. However, few re-ports of serious complications following neurovascular sheath can-nulation for surgical, radiological, or cardiac indications have beenreported. For example, during interventional cardiac procedures,large bore catheters are placed within brachial or femoral vessels.Heparin, LMWH, antiplatelet medications and/or thrombolytics aresubsequently administered. The small number of patients withhemorrhagic complications following peripheral or plexus blockis insufficient to make definitive recommendations. However,trends are evolving which may assist with patient management.For example, these cases suggest that significant blood loss, ratherthan neural deficits may be the most serious complication of non-neuraxial regional techniques in the anticoagulated patient. Inaddition, hemorrhagic complications following the deep plexus/peripheral techniques (lumbar sympathetic, lumbar plexus, andparavertebral), particularly in the presence of antithrombotic ther-apy, are often serious and a source of major patient morbidity.While needle/or catheter placement was described as difficult,

ulant drugs.

l procedure and evaluate INR. INR should be within the normal range at time ofnt factors. Postoperatively, daily INR assessment with catheter removal occurring

ridine derivatives (clopidogrel and ticlopidine) should be discontinued 7 ands should be discontinued to allow recovery of platelet function prior to procedure

n procedure and initiation or discontinuation of these medications. Follow

prophylaxis LMWH dose. For ‘‘treatment” dosing of LMWH, at least 24 h shouldd within 24 h after the procedure. Indwelling epidural catheters should beLMWH and strict avoidance of additional hemostasis-altering medications,

l daily dose is less than 10,000 units. The risk with higher dosing regimens is

ment normal aPTT. Heparin may be restarted 1 h following procedure. Sustainedted with increased risk; monitor neurologic status aggressively

Ib/IIIa inhibitors; INR, international normalized ratio; LMWH, low molecular weight

54 T.T. Horlocker / European Journal of Pain Supplements 3 (2009) 49–54

there is often no evidence of vessel trauma (including the patientdeath from massive bleeding).

In summary, the decision to perform spinal or epidural anesthe-sia/analgesia and the timing of catheter removal in a patientreceiving antithrombotic therapy should be made on an individualbasis, weighing the small, though definite risk of spinal hematomawith the benefits of regional anesthesia for a specific patient. Alter-native anesthetic and analgesic techniques exist for patients con-sidered an unacceptable risk. The patient’s coagulation statusshould be optimized at the time of spinal or epidural needle/cath-eter placement, and the level of anticoagulation must be carefullymonitored during the period of epidural catheterization (Table 3).Indwelling catheters should not be removed in the presence oftherapeutic anticoagulation, as this appears to significantly in-crease the risk of spinal hematoma. It must also be rememberedthat identification of risk factors and establishment of guidelineswill not completely eliminate the complication of spinal hema-toma. In Vandermeulen’s series, although 87% of patients had ahemostatic abnormality or difficulty with needle puncture, 13%had no identifiable risk factor (Vandermeulen et al., 1994). Vigi-lance in monitoring is critical to allow early evaluation of neuro-logic dysfunction and prompt intervention.

Conflict of interest statement

None.

References

Cheney FW, Domino KB, Caplan RA, Posner KL. Nerve injury associated withanesthesia: a closed claims analysis. Anesthesiology 1999;90:1062–9.

CLASP (Collaborative Low-dose Aspirin Study in Pregnancy) Collaborative Group.CLASP: a randomised trial of low-dose aspirin for the prevention and treatmentof pre-eclampsia among 9364 pregnant women. Lancet 1994;343:619–29.

Geerts WH, Pineo GF, Heit JA, Bergqvist D, Lassen MR, Colwell CW, et al. Preventionof venous thromboembolism: the seventh ACCP conference on antithromboticand thrombolytic therapy. Chest 2004;126:338S–400S.

Goldstein S, Dean D, Kim SJ, Cocozello K, Grofsik J, Silver P, et al. A survey of spinaland epidural techniques in adult cardiac surgery. J Cardiothorac Vasc Anesth2001;15:158–68.

Ho AM, Chung DC, Joynt GM. Neuraxial blockade and hematoma in cardiac surgery:estimating the risk of a rare adverse event that has not (yet) occurred. Chest2000;117:551–5.

Horlocker TT, Bajwa ZH, Ashraf Z, Khan S, Wilson JL, Sami N, et al. Risk assessment ofhemorrhagic complications associated with nonsteroidal antiinflammatorymedications in ambulatory pain clinic patients undergoing epidural steroidinjection. Anesth Analg 2002;95:1691–7.

Horlocker TT, Wedel DJ. Neuraxial block and low-molecular-weight heparin:balancing perioperative analgesia and thromboprophylaxis. Reg Anesth PainMed 1998;23:164–77.

Horlocker TT, Wedel DJ, Benzon H, Brown DL, Enneking FK, Heit JA, et al. Regionalanesthesia in the anticoagulated patient: defining the risks (the second ASRAconsensus conference on neuraxial anesthesia and anticoagulation). Reg AnesthPain Med 2003;28:172–97.

Horlocker TT, Wedel DJ, Schlichting JL. Postoperative epidural analgesia and oralanticoagulant therapy. Anesth Analg 1994;79:89–93.

Horlocker TT, Wedel DJ, Schroeder DR, Rose SH, Elliott BA, McGregor DG, et al.Preoperative antiplatelet therapy does not increase the risk of spinal hematomaassociated with regional anesthesia. Anesth Analg 1995;80:303–9.

Kovesi T, Royston D. Is there a bleeding problem with platelet-active drugs? Br JAnaesth 2002;88:159–63.

McEvoy MD, Bailey M, Taylor D, Del Schutte Jr H. Noncompliance in the inpatientadministration of enoxaparin in conjunction with epidural or spinal anesthesia.J Arthroplasty 2000;15:604–7.

Moen V, Dahlgren N, Irestedt L. Severe neurological complications after centralneuraxial blockades in Sweden 1990–1999. Anesthesiology 2004;101:950–9.

Odoom JA, Sih IL. Epidural analgesia and anticoagulant therapy. Experience withone thousand cases of continuous epidurals. Anaesthesia 1983;38:254–9.

Rao TL, El-Etr AA. Anticoagulation following placement of epidural andsubarachnoid catheters: an evaluation of neurologic sequelae. Anesthesiology1981;55:618–20.

Rose KD, Croissant PD, Parliament CF, Levin MB. Spontaneous spinal epiduralhematoma with associated platelet dysfunction from excessive garlic ingestion:a case report. Neurosurgery 1990;26:880–2.

Ruff RL, Dougherty Jr JH. Complications of lumbar puncture followed byanticoagulation. Stroke 1981;12:879–81.

Schroeder DR. Statistics: detecting a rare adverse drug reaction using spontaneousreports. Reg Anesth Pain Med 1998;23:183–9.

Schwander D, Bachmann F. Heparin and spinal or epidural anesthesia: decisionanalysis. Ann Fr Anesth Reanim 1991;10:284–96.

Tryba M. Epidural regional anesthesia and low molecular heparin: pro. AnasthesiolIntensivmed Notfallmed Schmerzther 1993;28:179–81.

Vandermeulen EP, Van Aken H, Vermylen J. Anticoagulants and spinal–epiduralanesthesia. Anesth Analg 1994;79:1165–77.

Wu CL, Perkins FM. Oral anticoagulant prophylaxis and epidural catheter removal.Reg Anesth 1996;21:517–24.