Spinal or epidural anesthesia, either alone or in conjunction with general anesthesia, has been purported to have several major advantages over general anesthesia alone. Depending on the surgical procedures, spinal or epidural anesthesia has been suggested to decrease blood loss, 1 decrease the rate of deep venous thrombosis, 1 and, when combined with postoperative pain control, decrease the rate of major cardiopulmonary complications. 2 One group studying peripheral vascular surgical patients found a lower incidence of graft thrombosis in the epidural group 3 along with an increase in markers for fibrinolysis, a possible mechanism for this result. 4
If these data, extolling the benefits of central neuraxial anesthesia, are believed, and spinal or epidural anesthesia has major benefits for patients, then these techniques should be used in caring for the sickest patients. Often the very sickest patients come to the operating suite for a procedure requiring some type of anticoagulation or have, in fact, already been given medication to thin their blood.
“Under what conditions is it safe to administer an epidural or spinal anesthetic in patients who are, or will be, anticoagulated?” is a question for which there is no clear-cut answer. In these patients, clinical practice must follow the available scientific data, good quality assurance practice, and the risk-benefit ratio for each individual patient. The purpose of this chapter is to review the major complication of regional anesthesia in anticoagulated patients: epidural hematomas with resultant permanent paralysis. By reviewing individual case reports in which epidural hematomas occurred, coupled with data from studies showing safety of regional anesthesia in anticoagulated patients, it is hoped that informed decisions may be made in caring for critically ill patients.
Studies Showing Safety
Large Case Reviews
Spinal and epidural anesthetics have been reviewed in large case series from the 1950s. In one of the landmark studies in anesthesia, Dripps and Vandam 5 reviewed 10,098 spinal anesthesia cases and found only 66 patients who complained of a peripheral neuropathy. No “central nervous system sequelae,” including cauda equina syndrome, was noted in any of these patients. Similarly, in 1975, work summarized by Usubiaga, 6 in which approximately 750,000 epidural anesthetic cases were reviewed, showed fewer than 100 minor complications like the peripheral neuropathy found in the Dripps and Vandam spinal study. In Usubiaga's review, however, eight epidural hematomas were reported.
Vandermeulen and others, 7 in their 1994 study “Anticoagulants and Spinal-Epidural Anesthesia,” listed 11 references of reviews of epidural anesthesia in which no epidural hematomas are reported. The patient samples in these 11 references total >258,000, and the studies date from 1957 to 1989. Similarly, these authors quote seven references, including the Dripps and Vandam review, showing the absence of epidural hematomas after spinal anesthesia. These reviews, dating from 1952 until 1979, contain >65,000 patients.
Finally, Vandermeulen and coworkers 7 discussed reviews of Tryba, 8 who reviewed seven case series of subarachnoid anesthesia in 1993 without noting an epidural hematoma. The number of patients in these series was reported at 650,000. Tyrba also comments on 13 case series of epidural anesthesia numbering 850,000 patients in which only three epidural hematomas were discovered. Vandermeulen and colleagues commented that Tryba did not specify whether his data came from new reviews and, as in most other reviews, did not specify whether the patients were taking anticoagulation or had procedures requiring anticoagulation.
From these case series, it is clear that central neuxial anesthesia can be quite safe for a large number of surgical patients. However, safety data in patients undergoing surgical procedures in which anticoagulation is either present preoperatively or used intraoperatively are not found in the large reviews cited previously. However, several studies and case reviews reported spinal or epidural anesthesia in patients who are undergoing or who underwent anticoagulation.
Spinal and Epidural Anesthesia in Vascular Surgical Patients
In 1981, Rao and El-Etr 9 reported on 4,015 patients who required lower extremity surgery and consented to a regional anesthetic. Of this group, 3,164 underwent an epidural (using, at that time, either 1.5% lidocaine, 1.5% mepivacaine, or 0.75% bupivacaine, with about half receiving epinephrine), and 847 were given a continous spinal with tetracaine. The authors were meticulous in their patient selection, in their handling of the catheter placement and removal, and in their tight control of the intraoperative heparinization scheme. Exclusion from regional anesthesia included “patients with a history of leukemia, hemophilia, blood dyscrasias, thrombocytopenia or preoperative anticoagulation therapy.” Catheters were placed with a midline approach with a 17-gauge Tuohy needle and threaded 2 cm in either the subarachnoid or epidural space. In four patients a bloody tape occurred; their procedures were canceled, and they were brought back after 24 hours and underwent a general anesthesia for their peripheral vascular surgery. All catheters were removed 24 hours after insertion and “1 hour prior to the administration of the maintenance dose of heparin.” Intraoperatively, the authors monitored heparin levels using an activated clotting time (ACT). The goal was to reach an ACT of twice baseline value. In these patients, the baseline ACT was 76 ± 28 seconds and, after an average heparin dose of 2,600 ± 400 units, ACTs were 174 ± 30 seconds. Heparin was given in incremental doses of 500 units every 3 minutes until the desired level was achieved and the total dose was repeated every 6 hours as needed.
It is important to review Rao and El-Etr's procedures carefully because one cannot argue with the safety of their data. In the 4,015 patients, none reported an epidural hematoma or significant permanent neurological complication. Among the epidural patients, 4 reported transient paresthesias postoperatively, whereas 1 of the 847 continuous spinal anesthesia patients reported this. Backache occurred in 9 patients in the epidural group and 6 in the spinal group. Because these complications were minor, this study has formed the classic article in the field regarding tight control of heparinization decisions whether to delay the procedure in the event of a traumatic tap.
Odoom and Sih, 10 in 1983, reported on 950 patients who underwent 1,000 similar peripheral vascular procedures. The lumbar epidural catheter was placed with an 18-gauge Tuohy needle using the loss or resistance to air technique in the lateral position after the induction of a general anesthesia. All patients came to the operating suite after administration of an oral anticoagulation medication. The dose of the medication was not specified but it was sufficient to give an abnormal “thrombotest” (reported as 19% compared with normal of 80–100%). Exclusions to a regional anesthetic included “neurological disease, infection at the puncture site, blood dyscrasias (including thrombocytopaenia, prior heparinization, aspirin therapy of long duration, and thrombotest below 10%)”. Heparin intraoperatively included a 2,000–4,000 unit load and an infusion at 250–300 μg/minute. Epidural catheters were removed 48 hours postoperatively.
Despite placing the epidural catheters in patients undergoing general anesthesia who had taken an oral anticoagulant, the 1,000 epidural patients reported here did not suffer from an epidural hematoma or from a transient minor neurological complication. This study seems to be the first in the literature in which spinal or epidural anesthesia was performed in patients who had been anticoagulated before the central neuraxis anesthestic.
Baron and colleagues 11 retrospectively reviewed their data of 912 vascular patients who underwent surgery under epidural anesthesia with lidocaine. Each patient was heparinized at a level of 75 U/kg (5,000 U/70 kg) and then placed on a 1,000-U/hour drip. All patients had a documented partial thromboplastin time (PTT) >100 seconds in the operating room. All epidural catheters were removed in the operating suite without waiting to check on the PTT, and in these 912 patients, no neurological event occurred.
Caudal Blocks in Anticoagulated Patients
A nice series of case reports of central neuraxial anesthesia in patients with abnormal coagulation status is that of Waldman and colleagues 12 from 1987. The authors discussed a series of 36 patients who were referred to their pain treatment center for therapy. Each patient had a medical condition in which their coagulation laboratory values were abnormal. All 37 patients had prothrombin time (PT) or PTT 1.5 times the control. Twelve of the 37 on intravenous heparin were medically judged not safe to stop this medication before the caudal block. Nineteen of the 37 had platelet counts <50,000. These patients were treated with a total of 336 caudal blocks. The blocks were performed while the patients were prone, a small (25 gauge) needle was used for entry into the caudal space and 20 mL of local anesthetic and narcotic was injected into the space. After the caudal, the patients were instructed to lie flat for 1 hour. Of the 336 caudal blocks in these completely anticoagulated patients, no epidural hematomas causing neurological dysfunction occurred. Only two patients had a small hematoma at the needle's insertion site.
Antiplatelet Therapy in Orthopedic Patients
Horlocker and colleagues, 13 in 1990, reported on 1,013 orthopedic patients who underwent 399 spinal anesthestics and 614 epidural anesthetics for their orthopedic procedures. In this group, 209 patients were on aspirin and 232 patients were on another antiplatelet antiinflammatory medication. No patients had an epidural hematoma despite being treated with these medications, but there was a small incidence of what the authors termed a “minor hemorrhagic complication” or a bloody tap in these groups. The aspirin group had a 7 of 209 incidence of a bloody tap, and the other nonsteroidal antiinflammatory group had 6 of 232 incidence of a bloody tap. Only 8 of 622 patients who were not on any antiplatelet drug had a bloody tap.
Spinal and Epidural Manipulation in Cardiac Surgical Patients
The earliest report of subarachnoid or epidural manipulation of cardiac surgical patients occurred in 1980, when Mathews and Abrams 14 reported on 40 patients who underwent a single-shot spinal anesthesia with 10 mg of morphine sulfate in the lumbar area. Since that time, there have been a number of reports of attempts at either intrathecal opioids or postoperative epidural analgesia in cardiac surgical patients. These patients, and reports, for the most part involve clinical scenarios in which the patients are fully heparinized for cardiolpulmonary bypass. Instead of the 2,600–5,000 units of heparin given for vascular surgical patients, the cardiac surgical reports discuss patients who are given 300–400 units/kg of heparin (21,000 units for a 70-kg patient). Most of these studies attempt to examine stress hormones, pulmonary function, and postoperative analgesia with mixed results. In 1997, Chaney 15 reviewed the literature for cardiac surgical patients and found 15 studies of intrathecal opioids in which 771 patients were involved and 13 studies consisting of 417 patients where epidural analgesia was studied. In none of these reports was an epidural hematoma implicated in a postoperative complication. Parenthetically, it should be added that most of the epidurals were placed the night before or as far before the heparinization as possible. Additionally, catheters were removed postoperatively after the coagulation status had normalized. As minimally invasive cardiac surgery develops, with decreased heparinization, expect more reports of these techniques in cardiac surgical patients to appear in the literature.
Case Reports of Epidural Hematomas
Even though large case series of spinal or epidural anesthesia in patients who are or are not anticoagulated indicate that an epidural hematoma is an infrequent occurrence, the fact still remains that there are many reported cases of this complication in the literature. When this complication occurs, it is devastating for the individual patient and represents a nightmare for the regional anesthetist. All cases discussed next involve patients who undergo a spinal or epidural anesthetic for a surgical procedure and become paralyzed. Typically, the syndrome begins with severe back pain and then progresses to flaccid paralysis. Sensory disturbances usually occur after the motor disturbance and a large mass compressing the spinal nerves is confirmed by surgery, computed tomography scan, magnetic resonance imaging, or autopsy.
In 1986 Owens and others 16 published the first systematic review of epidural hematomas after spinal anesthesia or lumbar puncture. The authors discussed the history of a 70-year-old man who underwent a hernia repair and, on Postoperative Day 1, suffered what was most likely an embolic stroke. As part of the work-up for this neurological dysfunction, the patient underwent a lumbar puncture that was “technically difficult” at 5 pm and was heparinized with a 5,000-unit bolus followed by a 1,000-unit/hour infusion over 2 hours later. PTT was greater than 120 seconds, and by the next morning, after a night of severe progressive back pain, the patient was noted to have bilateral lower extremity paralysis and decreased sensation to pinprick below L-1. The patient underwent rapid evacuation of a clot but did not regain use of his lower extremities.
From this index case report, the authors reviewed the literature on epidural hematomas after spinal anesthesia and after lumbar puncture. They found 33 prior cases of subarachnoid manipulation followed by a hematoma dating from 1911 to 1981. Of these 34 total cases, 28 had an anticoagulant coagulopathy, according to the authors. This included 9 patients with systemic cancer with a platelet count of <44,000; 14 who had work-up for a stroke or a transient ischemic attack and had received heparin (n = 7), coumadin (n = 1), or both (n = 6); and 5 patients with a coagulation risk factor and who were reported to have cirrhosis of the liver, an abnormal bleeding time, an abnormal PT, or aspirin treatment or had received the then-new antiplatelet drug ticlodipine. Of the 6 patients who did not have a hematological risk factor, 3 were treated for central nervous system infection and one had no coagulation labs but had cirrhosis of the liver on autopsy. In fact, only 2 had no hematological abnormality.
Only six of the patients in Owens and colleagues 16 report experienced an epidural hematoma after a spinal anesthetic. Four of the six had a documented bloody tap and the cases of the other two were described as technically difficult. One case involved a continuous spinal anesthesia with a 16-gauge needle that was intended to be a epidural but, after the wet tap, was managed this way. Two of the six patients were treated with antiplatelet medications (one with aspirin postoperatively and one with ticlodipine preoperatively). Two of the patients indeed had no prior hematological risk factor and their epidural hematoma was attributed to a difficult placement.
In the 1975 International Anesthesia Clinics Jose Usubiaga 6 presented the first report of epidural hematomas after epidural anesthesia. Usubiaga, an Argentinian anesthetist who retrained in the United States, was an early proponent of epidural anesthesia. In his report he reviewed case series throughout the world on epidural anesthesia and found eight patients with epidural hematomas among an estimated 750,000 patients. Four of these 8 has been treated postepidural with intravenous heparin and 3 had a traumatic, bloody tap; only a single patient was without either of these risk factors.
The next, most current review of epidural hematomas after spinal or general anesthesia was published in 1994 by Vandermeulen and coworkers. 7 In this review, compiled from cases from the National Library of Medicine MEDLINE system and by “previously published journal reviews and large case series,” 61 published cases were found of an epidural hematoma after a spinal or epidural anesthetic. Of these 61 cases, 46 followed an epidural anesthestic and 15 followed a spinal anesthestic. Of the 61 patients, 42 had “impaired coagulation” sometime during their therapy. Heparin was seen in 30 of 42, and the other 12 cases had other “medical conditions” such as aspirin, ticlodipine or urokinase therapy along with history of thrombocytopenia or alcohol abuse. Of the 19 cases in which there were no clotting risk factors, 4 were traumatic, 3 involved spinal tumors, and 1 was an omphalocele repair in a very ill neonatal patient.
The authors stated that 15 of the 61 epidural hematomas occurred after a bloody tap and another 15 occurred after a difficult insertion. Nearly 87% (53/61) epidural hematomas occurred after either a bloody insertion or a traumatic insertion or when the patient had some type of anticoagulation; 20 of the 61 reported epidural hematomas had both a difficult, bloody tap and were treated with an anticoagulant.
This review offers insight into the signs and symptoms of an epidural hematoma, the time course of onset of symptoms, and the importance of rapid decompression. In 15 patients, the onset of the hematoma occurred after removal of a catheter. Muscle weakness was the first sign in 28 of 61 patients, back pain in 23, and sensory deficit in 9. Paraplegia was recorded to occur within 14.5 ± 3.7 hours after ending of the anesthetic. Surgical decompression occurred in only 38 patients with epidural hematomas, and the most successful recovery occurred in patients in whom surgical intervention was performed in less than 8 hours.
As of 1994, patients in whom an epidural hematoma developed after a spinal or epidural anesthetic could be expected to have a bloody, difficult tap or to have had perioperative anticoagulant treatment. Vandermeulen and colleagues reported only 3 patients who had been exposed to fractionated or low-molecular-weight heparin (LMWH) in whom an epidural hematoma developed. These data, however, were to change.
The Low-Molecular-Weight Heparin Story
Standard heparin is a heterogenous long-chained sugar molecule containing 40–50 sugar moieties with a molecular weight of 12,000–15,000. It has a high protein-binding property (and hence is difficult to predict how an individual patient will react). Standard heparin acts mainly by binding and activating a serum protein molecule called antithrombin III such that the heparin-antithrombin III complex binds several serine proteases in the coagulation cascade. Although this complex has the strongest avidity for thrombin, it also binds the factors IXa, Xa, XIa, and XIIa. Standard heparin also has strong platelet inhibitory properties. This heparin will elevate standard laboratory tests of the PTT and PT. 17
Fractionated or LMWH is either chemically or enzymatically cleaved such that the molecular weight of this substance is from 4000–6500 d with from 13–22 sugar units. LMWH has little protein binding, is mainly cleared by the kidney, and has very little antithrombin III activity. Instead, it has a very active factor Xa binding and does not increase the standard tests (PTT) of heparin activity. It has shown to be more effective in prophylaxis for deep venous thrombosis than any other anticoagulation treatment for total knee replacement and as efficacious as in prophylaxis for total hip surgery and major trauma. 17
LMWH rapidly became associated with case reports of epidural hematomas. In the last systematic review of these reports in 1998, 18,19 52 cases of epidural hematomas after spinal or epidural anesthesia have been associated with administration of LMWH. Twelve have been reported in Europe, 19 and 40 have been reported with the U.S. Food and Drug Administration Med Watch program. 18
Of these 52 cases, 34 have been from epidural placements, 14 from spinal attempts, and 4 from a source not specified in the report. Of the 12 patients from Europe, 6 received other anticoagulants that impact platelet function (aspirin, ketorolac) and 5 were noted to have involved difficult placements or bloody placements.
Of the 40 epidural hematomas from the United States, 26 were from epidurals, 10 from spinal anesthetics, and 4 from unidentified sources. Of the 10 spinal anesthesia patients, 3 received an antiplatelet medication, 1 received intravenous heparin in addition to the LMWH, and 6 received the LMWH on the operative day. Of the 26 epidural-associated hematomas in the United States, 12 were also associated with other anticoagulation. In 10 of the 26, the time of onset of symptoms was less than 12 hours after removal of the epidural catheter.
American Society of Regional Anesthesia Consensus Panel on Regional Anesthesia and Anticoagulation
Given the controversy over LMWH and the lack of solid randomized, prospective data, the American Society of Regional Anesthesia (ASRA) convened a panel of 17 experts in the field to analyze available data and to present a consensus of action based on the data. The panel commented on spinal and epidural anesthesia not only on LMWH but in four other areas: oral anticoagulants, antiplatelet drugs, fibrinolytic and thrombolytic drugs, and standard heparin. The standard recommendation common for all areas was to “monitor neuro status carefully.” Given the clear data that the best chance for recovery of neurological function after an epidural hematoma comes from rapid decompression, the importance of this recommendation is clear.
For these medications the ASRA panel had seven recommendations. 20 First, the panel was emphatic in stating that one should stop the coumadin and measure a normal PT/INR before starting the regional anesthetic in patients on this medication. Second, they urged caution when other anticoagulant medications, such as heparin, or antiinflammatory agents are used in conjunction with coumadin when considering a regional technique. Third, it urged checking the PT and International Normalized Ratio (INR) if the first dose of prophylactic coumadin for the surgical procedure is after 24 hours, and, fourth, to check the PT/INR daily if the patient is on more than low doses of coumadin (5 mg/day). Fifth, markedly elevated PT or INR levels may alert one to a patient who is sensitive to the medication, and consideration to decreasing the coumadin dose should be entertained before removing an epidural catheter, although the panel made no definitive recommendations on removing catheters in patients with “therapeutic levels of anticoagulation” other than to “exercise clinical judgment.” Last, elderly, ill (renal, hepatic, cardiac diseases) women are at higher risk for sensitivity to coumadin, and the dose may need to be adjusted.
The ASRA emphatically stated that the risk of epidural hematoma was very low if antiplatelet drugs were used alone without other blood-thinning medications. 21 They also stated that there was no apparent concern for risk in use of the nonsteroidal antiinflammatory medications, and that there were no data on combination of antiplatelet medications. However, they cautioned on the potential for epidural hematoma when these medications are used in conjunction with LMWH. Finally, the committee indicated that there was no acceptable single test that was useful in determining the extent of platelet function (including bleeding time).
The ASRA consensus panel indicated that patients who received fibrinolytics were at extremely high risk for “adverse neuraxial bleeding” if a block is attempted, and that these types of anesthetics should be administered only in “unusual circumstances.”22 There were no consistent data on the individual duration of action of the fibrinolytics; therefore, the task force did not offer any recommendations on the time for catheter removal other than the fact that careful neurological monitoring was essential.
The ASRA Committee indicated that standard subcutaneous heparin alone did not pose any increased risk of epidural hematomas if the patient was not debilitated or had been on the subcutaneous heparin for a prolonged time. In intravenous standard unfractionated heparin infusion, the ASRA recommended that it could be used safely for vascular surgery if its administration was delayed for 1 hour after needle placement and any catheter removed 1 hour before a subsequent heparin dose or 2–4 hours subsequent to the last heparin dose. The committee did not comment on the safety of central neuraxial blockade in cardiac surgical patients, stating that the data were inconclusive. 23
It was in this area that the ASRA Consensus Panel had the most to say. 18,19 Clearly, use of LMWH in combination with another antiplatelet medication increased the risk of epidural hematoma. If blood appears in a needle during the block, the committee recommended that the LMWH dosage be delayed for 24 hours. In patients already on LMWH on arrival to the operating room, it was thought that at least 12 hours and preferably longer should elapse before attempting a spinal or epidural anesthetic. The group also recommended the time interval between LMWH dose and catheter manipulation be as long as possible. Catheter removal “should be delayed for at least 10–12 hours after a dose of LMWH and further LMWH doses should be given at least 2 hours after any catheter is removed. Finally, the ASRA did not recommend the routine measurement of activated factor X, which in many hospitals can be done within an hour, indicating that the test is not a reliable measure of absence of LMWH activity.
The preceding discussion on regional anesthesia, specifically spinal and epidural anesthesia and anticoagulation, was presented as a guideline for readers. The key to successfully treating these patients with a spinal or epidural anesthetic is to remember that the benefits of a regional anesthesia should be considered relative to the risks in each individual case. It is hoped that this discussion will arm readers with data on the low risk of an epidural hematoma and guide them in considerations in their specific practices.