Portal de Periódicos da CAPES

Safety Steps for Epidural Injection of Local Anesthetics

1997; Lippincott Williams & Wilkins; Volume: 85; Issue: 6 Linguagem: Inglês

10.1097/00000539-199712000-00030

1526-7598

Michael F. Mulroy, Mark C. Norris, Spencer S. Liu,

Pain Management and Opioid Use

Epidural postoperative analgesia is a popular technique that may improve patient comfort and outcome while reducing hospital costs [1,2]. However, significant risks accompany the injection of local anesthetic into the epidural space, including unintentional subdural, subarachnoid, and intravenous (IV) drug injection. Despite years of use, controversy still surrounds the appropriate steps to ensure the safe initiation of epidural anesthesia. Risks Associated with Epidural Blockade Subdural Injection Subdural injection is the least common misadventure. The subdural space is difficult to enter deliberately, because the arachnoid is closely adherent to the overlying dura in most patients. Entry into this space may not be recognized because cerebral spinal fluid will not flow freely, either spontaneously or with aspiration. Injection of the local anesthetic solution into this small space produces a high level of sensory and motor anesthesia, but with a slower onset than a subarachnoid injection. Subdural anesthesia resembles epidural anesthesia in its speed of onset, although the duration is usually shorter. This clinical picture is suggestive, but definitive confirmation can only be obtained by injection of radiocontrast dye [3-7]. A retrospective review of a pain clinic experience reports an incidence of 0.82% subdural injections, based on a clinical picture of unexpected high "epidural" anesthesia [8], but this report may overestimate the true frequency because of the patient population and the criteria used. Despite several case reports of subdural injection, it seems to be a relatively rare occurrence, as it is not mentioned as a complication of epidural techniques in large series. Intrathecal Injection Accidental dural puncture ("wet tap") is a more common complication. The reported frequency varies between 0.61% and 10.9% (Table 1) [9-16]. Because it is usually readily recognized, a "total spinal" from unintentional injection is relatively rare. Subarachnoid placement is less frequent but more difficult to appreciate when it occurs as a result of migration of the tip of the epidural catheter, during either initial insertion [17] or the course of ongoing epidural therapy, with an incidence reported between 0.26% [11] and 0.6% [16].Table 1: Incidence of Dural Puncture and Total Spinal During Epidural AnesthesiaIntravascular Injection The most significant hazard of epidural blockade is unrecognized unintentional intravascular injection. This is more likely with epidural anesthesia than with other regional techniques because of the number of venous plexuses in the epidural space, and it may be at least partially related to the relatively low pressure existing in these veins. This low pressure, usually equated to intrathoracic pressure, may produce neither a spontaneous flow of blood from a punctured vein nor a positive flow with aspiration; the veins collapse when further negative pressure is applied. Vessel entry can occur on initial insertion of the epidural needle or on insertion of the catheter, or, again, because of migration of the catheter tip at any time during ongoing epidural therapy [18,19]. The frequency of vessel entry ranges from 0.2% to 11% (Table 2) [9,10,13,20-27]. Although the typical frequency seems to be in the range of 2%, vessel entry may occur more often in obstetrical patients [7%-8.5% incidence [28,29]] and in situations in which the epidural catheter is inserted before injection of the local anesthetic [up to 9% incidence [21,30]].Table 2: Frequency of Intravascular Injection with Epidural AnesthesiaIntravascular placement presents significant patient risks for two reasons. First, vessel entry is sometimes difficult to diagnose. Second, bolus injection of large volumes of local anesthetic can rapidly produce toxic drug concentrations. Seizures, respiratory arrest, and, with bupivacaine especially, cardiac arrest and death [31,32] can follow. Reports of bupivacaine-induced cardiac arrest in 1979 stimulated the current interest in test procedures to detect IV placement. Before this time, recommendations regarding safety steps were vague [33,34] and nonspecific [35,36]. Since 1979, there has been extensive interest in developing a simple and reliable method of detecting incorrect, especially intravascular, placement of an epidural needle or catheter. Desirable Properties of Ideal Test The ideal epidural "test" should be safe, simple to use, and incorporate readily available materials and monitoring devices. It must also reliably provide for detection of both intravascular and subarachnoid placement of drug and do so within a short time with minimal patient cooperation. It should have a highly sensitive and specific end point. It should be applicable under multiple situations, including the extremes of age and the presence of concomitant medications, including general anesthetics. Last, it should be safe so as to represent no additional risk to the patient when producing a positive result. Multiple regimens for testing epidural needles and catheters have been suggested, although none are ideal. Recommendations include aspiration, incremental injection, and the injection of markers of intravascular placement such as local anesthetics, epinephrine, air, ephedrine, isoproterenol, succinylcholine, and fentanyl. Critical Review of Current Components of Suggested Tests Aspiration Aspiration of the needle or catheter often allows the identification of subarachnoid and intravascular placement, but there are multiple reports of intravascular injection after negative aspiration [16,20,25] (Table 2). The ability of aspiration to detect intravascular catheter location may depend on the type of epidural catheter used. Aspiration has failed to detect intravascular placement when catheters with a single end-hole are used [25,37]. Some claim that aspiration always detects IV placement of multiholed catheters [38,39]. Systematic clinical trials have not addressed this issue. Aspiration is a useful first step but cannot be relied on as the sole test. Incremental Injection Incremental injection is the mainstay of safe regional anesthesia practice. Theoretically, if small injections are used, the patient will report subjective symptoms before blood local anesthetic concentration reaches the threshold for seizures [25,40]. Even if subjective symptoms are not reported, the use of increments might reduce the peak blood levels obtained before the first signs of seizure activity. Although this recommendation is a logical and appropriate safety step, the efficacy of incremental injection as a true test dose has not been substantiated in a large clinical series. One report of obstetrical patients included 12 of 51 women who received relatively large doses of local anesthetic without any systemic symptoms (the other 39 had seizures), despite subsequent aspiration of blood from the catheter [25], implying that incremental injection will not detect all intravascular injections. Local Anesthetic Test Injections Subarachnoid Injection. A 3-mL local anesthetic test dose is traditionally used to detect subarachnoid injection. Abraham et al. [20] reported that the onset of sensory anesthesia in the S2 dermatome occurs within 2 min when 3 mL of 1.5% hyperbaric lidocaine is injected in the subarachnoid space at the second lumbar level. In contrast, at least 3 min will pass before epidural injection will produce sensory changes, first seen in the L2 distribution. Subarachnoid isobaric bupivacaine has also been studied, but the variability of spread and the delay in onset with doses of 8-15 mg make this an inadequate test [41-46]. The perception of warmth in the feet has been suggested as a potential sign of subarachnoid injection but is not a reliable indicator [47]. If subarachnoid injection of a hyperbaric lidocaine test solution is suspected, a minimum of 2 min should pass before testing for sensory changes in the sacral area. With bupivacaine, a longer delay and a high index of suspicion, seeking multiple signs and symptoms that suggest intrathecal drug injection, are needed. There are risks associated with subarachnoid injection of test doses. In laboring women, significant hypotension and midthoracic (or higher) levels of sensory blockade can follow an intrathecal injection of 3 mL 1.5% lidocaine with 1:200,000 epinephrine [11]. Spinal anesthesia requiring intubation and ventilation has been reported after 3 mL 1.5% lidocaine (45 mg) [48], 2.5 mL 2% 2-chloroprocaine (50 mg) [49], and 3.0 mL 0.5% bupivacaine (15 mg) [42]. Intravascular Injection. Assessing for symptoms of central nervous system irritability produced by subtoxic doses of local anesthetic has also been recommended as an indicator of unintentional intravascular injection [35,36,50-52]. Symptoms include tinnitus, perioral tingling, metallic taste, dizziness, sedation, change in hearing, and multiple other subjective symptoms. Patients are often asked whether they can identify any of these signs after the injection of the standard 3-mL test dose. Unfortunately, there are several limitations of this practice. The amount of local anesthetic required to produce these symptoms is unclear but is probably greater than the total milligram dose usually contained in the standard 3-mL initial test. Lidocaine or chloroprocaine 100 mg [51] or 1 mg/kg [52] or 25 mg of bupivacaine [53] produces subjective symptoms in the unmedicated patient. This requires a minimum of a 5-mL volume (of 2% lidocaine or 0.5% bupivacaine) [40,44]. Preliminary data suggest that the dose of ropivacaine required to produce symptoms is 25% greater than that of bupivacaine [54]. A smaller test dose (to exclude subarachnoid placement) should precede the larger dose designed to identify intravascular placement. Unfortunately, local anesthetics do not produce symptoms in all clinical situations. Laboring women may not report dizziness, tinnitus, or metallic taste after chloroprocaine [51]. Premedication with sedatives such as the benzodiazepines may reduce the sensitivity of this test. Mulroy and colleagues [53] reported that 40% of patients premedicated with midazolam and fentanyl were unable to detect the presence of intravascular doses of chloroprocaine or bupivacaine that were otherwise easily appreciated by unmedicated volunteers. Although there are no large prospective series that quantify the efficacy of this test, it seems that local anesthetics might be an effective indicator of intravascular injection if used in appropriate doses and in unpremedicated patients. Epinephrine Epinephrine is the most extensively studied marker of IV injection (Table 3). Moore and Batra [55] initially reported the use of 15 micro g of epinephrine with local anesthetic in a series of 175 surgical patients. They observed an average increase in heart rate from 79 bpm to 110 bpm within 60 s of injection of the epinephrine test dose. The tachycardia resolved within 60 s. Subsequent studies have established that for unmedicated surgical patients, an increase of heart rate of 20 bpm and an increase of systolic blood pressure (SBP) of 15 mm Hg are sensitive indicators of intravascular injection [56]. Although no large prospective study of the efficacy of the epinephrine test dose has been published, it has become widely used since 1981.Table 3: Summary of Prospective Studies of Efficacy of Epinephrine Epidural Test Doses for Detection of Intravascular InjectionLimitations of Epinephrine. Although the epinephrine test dose meets the desired criteria of simplicity, use of available materials, easy monitoring, and clear end points, both patient conditions and concomitant medications limit its usefulness or require modification of the criteria for a positive test (Table 4).Table 4: Use of Epinephrine Test DosesAcute and chronic use of selective or nonselective beta adrenergic blocking agents reduce the pulse rate changes after epinephrine [56,57]. Although beta blockade attenuates the heart rate changes produced by IV epinephrine, an increase of systolic blood pressure of 15 mm Hg within 2 min is a reliable indicator of intravascular injection in the presence of beta blockade [56]. Studies of other cardiac medications, such as calcium channel blockers or antiarrhythmics, have not been reported. At 1-minimum alveolar anesthetic concentration (MAC) levels in children and adults, halothane [58] and isoflurane [59-61] attenuate the intrinsic hemodynamic variability and interfere with the tachycardic response to epinephrine. As a result, the magnitude of positive hemodynamic criteria should be adjusted in patients receiving potent inhaled anesthetics. During anesthesia with 1 MAC volatile anesthetics, an increase in heart rate of >8 bpm or of SBP >13 mm Hg is an appropriate indicator of intravascular injection of epinephrine. Regional anesthesia alone (in the form of spinal or epidural blockade) does not interfere with the heart rate response, but it can attenuate the increase in SBP from IV injection of epinephrine [62,63]. A high (T5) epidural combined with general anesthesia attenuates both the heart rate and the blood pressure response to an epinephrine test [62]. The chronotropic response to beta-adrenergic stimulation declines progressively with increasing age. Five times the dose of isoproterenol is required to produce a 25-bpm increase in heart rate at age 70 yr as at age 20 yr [64]. This same phenomenon occurs with epinephrine. The original report of Moore and Batra [55] did not include patients over the age of 60 yr. The question is unanswered whether the epinephrine test dose is adequate in the elderly. Two elderly patients have failed to respond to the epinephrine-containing test doses [65]. Shoenwald and colleagues [66] have shown a decrease in sensitivity of the epinephrine test dose in a group of vascular surgery patients (average age 67 yr), in whom a threshold of a 9-bpm increase in heart rate had to be used as a positive criterion. Guinard et al. [67] have shown a progressive decline with age in the heart rate increase in response to a (nonstandard) 10-micro g injection of epinephrine. In their series of 33 patients, the only false-negative 15-micro g test occurred in one patient over the age of 60 yr. These data suggest caution in interpretation of the epinephrine test dose in the elderly. Significant controversy surrounds the use of epinephrine-containing test doses in parturients. Although the epinephrine test dose has been recommended in obstetric practice [68], several factors alter the response to epinephrine in the parturient, and there is concern about the safety of intravascular epinephrine in the healthy and the preeclamptic patient. Limitations of Epinephrine in Parturients: Efficacy. Pregnancy alters a woman's sensitivity to chronotropes and vasopressors. Parturients require five times as much isoproterenol to induce a 25-bpm increase in heart rate as nonpregnant patients [69]. Unlike nonpregnant patients, the average parturient's heart rate increases by only 10 bpm after IV injection of the standard 15-micro g dose of epinephrine [70], and the average response is a brief tachycardia followed by bradycardia. Some women develop only a bradycardia [71]. On the other hand, women with preeclampsia have a substantial increase in sympathetic vasoconstrictor activity [72] and are much more sensitive than normal parturients to the chronotropic effects of isoproterenol and other vasopressors [73,74]. Another problem with the use of epinephrine in parturients is the presence of confounding heart rate changes during labor. Women often develop a significant tachycardia with each painful uterine contraction; distinguishing between an increase in heart rate caused by IV epinephrine and an increase caused by pain is difficult. More than 25% of laboring women have heart rate increases of at least 20 bpm during the induction of epidural analgesia without injection of an epinephrine test dose [75]. Injecting the epinephrine test dose between uterine contractions may limit this confusion. In active labor, however, contractions can arise so often that one unavoidably begins injection within 60 s of the test dose injection. In one study of actively laboring women, 15 of 20 participants began a contraction within 60 s of the study solution injection [71]. In another study, the epinephrine test dose had to be repeated in 13 of 206 laboring women because of coincident uterine contractions [28]. In clinical practice, as well, the epinephrine test dose can be unreliable. In a teaching center, 827 parturients had labor epidural analgesia induced with 1.5% lidocaine with epinephrine 1:200,000. Approximately 4% of the epidural catheters proved to be IV in location. However, despite the routine use of epinephrine, 0.6% (5 of 827) women had local anesthetic-induced seizures.1 (1) Naulty JS, March MG, Leavitt KL, et al. Effects of changes in labor analgesic practice on the safety and efficacy of epidural anesthesia [abstract]. Anesthesiology 1992;77:A983. Limitations of Epinephrine in Parturients: Safety. IV epinephrine causes a significant increase in blood pressure [71]. Women with preeclampsia and increased blood pressures may respond to IV epinephrine 15 micro g with dangerous hypertension. Even epidural epinephrine can provoke a hypertensive crisis in a parturient with an increased blood pressure [76]. In gravid ewes and guinea pigs, test doses of epinephrine decrease uterine blood flow by 50% for 2-5 min [77,78]. In a clinical study, the fetuses of 2 of 10 laboring women developed new late decelerations for up to 10 min after maternal IV epinephrine 15 micro g, although there was no adverse clinical outcome [71]. Despite these negative aspects, others feel that epinephrine 15 micro g can be a useful marker of IV injection in laboring women. Colanna-Romano et al. [70] reported that epinephrine 15 micro g does reliably increase maternal heart rate at least 10 bpm. In addition, they suggest that the tachycardia induced by epinephrine 15 micro g can be distinguished from pain-induced increases in heart rate by its faster onset [79]. In a prospective study of 209 laboring women by the same authors, using a 100-mg lidocaine injection as a confirmatory test, epinephrine 15 micro g detected all 14 intravascular epidural catheter placements [28]. Although this study is the most enthusiastic in supporting the use of epinephrine 15 micro g in the clinical setting, limitations were noted: there were contraindications to epinephrine use in 10 of the 209 patients studied-3 had equivocal responses, epinephrine injection was repeated in 13 women because a uterine contraction began shortly after the first epinephrine injection, and 8 false-positive responses (4.5%) led to unnecessary epidural catheter replacement. In summary, the epinephrine 15-micro g test dose has significant limitations as a marker of intravascular injection in laboring women: it is at times insensitive and often nonspecific, and the issue of fetal safety remains unresolved [80]. Safety of Epinephrine Test Doses. In addition to these limitations, there have been concerns about the safety of epinephrine-containing test doses in nonpregnant patients. The increases in heart rate and blood pressure associated with the IV injection of epinephrine increases myocardial oxygen consumption and could predispose a high-risk subgroup of patients to myocardial ischemia or cardiac arrhythmia [81]. The original report by Moore and Batra [55] included 95 patients who developed electrocardiogram changes in lead II during the time of the test dose injections, including one case of bigeminy. On the other hand, there have been no reports in the anesthesia literature of patients suffering an adverse outcome from a positive epinephrine test dose. Nevertheless, this concern may justify the use of alternative test doses and the reliance on incremental injection in those patients who are identified to be at significant risk of myocardial ischemia if an epinephrine-containing test dose were injected IV. Alternative Test Doses In view of the limitations of epinephrine, especially in pregnancy, several other markers of intravascular injection have been recommended and investigated. Air. In 1989, Leighton and Gross reported that air 1-2 mL was an effective marker of intravascular injection [82]. By placing the fetal external heart rate probe over the mother's sternum, changes in heart sounds induced by small amounts of IV air are readily detected. The air-Doppler test is easy to perform in the labor suite, is independent of maternal heart rate changes, and has no fetal effects. In a series of 313 patients, it was very sensitive (false-negative rate 0%, 95% confidence limits 0.0%-1.1%) and specific (false-positive rate 2%, 95% confidence limits 0.7%-4.3%) for intravascular injection [37]. It has been effective in a case report [83], and there are no reports of failures with this test. The safety of the air injection test in routine clinical practice has yet to be determined. IV entrainment of air has been reported frequently in obstetrical practice [84] with rare sequelae, which suggests that air test problems are unlikely. A greater limitation is that the Doppler monitor may not be readily available in every operating room, and this factor may limit the application of the air-Doppler test to the obstetrical suite. Isoproterenol. Another beta-agonist drug, isoproterenol, has also been advocated as a potential alternative to epinephrine, especially in adults and children receiving general anesthesia [85-87], in whom 0.1 micro g/kg or a total dose of 3 micro g will produce a heart rate increase of 20 bpm (Table 5). In a series of 10 obstetrical patients, IV isoproterenol produced a reliable increase in the heart rate [88]. In sheep, it does not affect uterine blood flow [89]. There are limitations to the potential use of isoproterenol in the parturient. First, its safety on intravascular injection (especially in the pregnant patient) has not been substantiated. Second, sensitivity to this drug may be reduced in pregnancy [69]. Third, there are no published neurohistological data documenting the safety of epidural or subarachnoid injection of isoproterenol.Table 5: Summary of Prospective Studies of Alternative Epidural Test Doses for Detection of Intravascular InjectionEphedrine. Fong [90] found that ephedrine increased maternal blood pressure but that the changes could not be distinguished from pain-induced hypertension. Cherala et al. [91] required a complicated calculation of the peak-to-peak change in heart rate frequency to detect the presence of the intravascular ephedrine. This complexity and lack of specificity make it unlikely that ephedrine will assume a practical role as an IV marker. Fentanyl. Fentanyl has been advocated as a potential marker of unrecognized intravascular injection [92] and has the advantages that it is commonly available, nontoxic to the spinal cord, and has no effects on uterine blood flow. Among laboring women, 92% correctly identify the route of injection (IV or epidural) of 100 micro g fentanyl [93], but some patients report dizziness and drowsiness even after epidural fentanyl [94]. Some do not report any symptoms after IV fentanyl [95]. Succinylcholine. Succinylcholine may be useful in producing transient muscle weakness under general anesthesia (in the absence of other muscle relaxants), when epinephrine may not be reliable [24,96], but its limitations in the awake patient are obvious. Efficacy and Limitations of Safety Steps Despite extensive awareness of the risks of epidural anesthesia, there remains a small but apparently unavoidable risk of misadventure. Specifically, subdural and subarachnoid injections will happen, although careful attention to testing and appropriate dosing seems to make serious side effects rare with these two complications. IV insertion of a needle or catheter also seems unavoidable. The historical risk varies between 0.2% and 11%, with a higher frequency in parturients and with insertion of epidural catheters. In contrast to the relatively stable rate of subdural and subarachnoid injections with epidural anesthesia, there seems to have been a decline in the frequency of systemic toxic reactions with this technique since approximately 1980. Although several studies previously reported high incidences of intravascular injection (Table 2), recent reports of large series of patients with epidural techniques for postoperative analgesia fail to mention a single case of systemic toxic reaction as a complication of epidural insertion [97-99]. In surgical use of epidurals, Tanaka et al. [10] reported an incidence of 0.12% of systemic toxic reactions in 17,000 epidurals performed in a teaching hospital with the use of aspiration and local anesthetics without epinephrine as a test dose before injection. In a series of 16,000 epidurals performed with epinephrine-containing test doses, Brown and colleagues [23] encountered 2 patients who experienced systemic toxic reactions to local anesthetics. Selander et al. [100] reported 5 patients who experienced systemic symptoms of local anesthetic toxicity (but no convulsions) associated with epidural injections during the preclinical trials of ropivacaine in approximately 3000 patients. In obstetrical practice, Hawkins et al. [101] reported "an abrupt decrease in deaths due to regional anesthesia" since 1984 compared with the frequency of deaths primarily related to local anesthetic toxicity in the previous 6 yr (based on data from the Pregnancy Mortality Surveillance information from the Center for Disease Control). These data do not support any conclusions about the cause in the reduction of mortality, although Naulty et al.'s report1 suggests that the decline in systemic toxicity is related to a shift to lower concentrations of local anesthetics, smaller initial doses, and the use of infusions rather than boluses. They reported a decrease in the frequency of systemic toxicity from 0.6% to 0.02% when those changes were made in 1987 in their practice (Lennart Waldenlind, Astra PC, Sweden, personal communication, 1997). A review of the reports of adverse events submitted to the Astra Pharmaceutical Company also shows an apparent decline in the frequency of systemic toxic reactions. Between 1984 and 1988, 33 systemic toxic reactions and 98 adverse reactions to bupivacaine were reported, in contrast to 13 and 36, respectively, from 1991 to 1995, whereas total local anesthetic sales increased threefold in that time. Unfortunately, the validity of these data is unclear: they are self-reported and do not indicate the true incidence or the denominator of the frequency of these events. If there has been a decline in the true frequency of systemic toxic reactions, it is unclear whether this is due to: (a) an enhanced awareness of this risk, (b) the increased use of incremental injection, (c) the adoption of epinephrine-containing (and other) test doses, (d) the use of lower concentrations and smaller total doses of local anesthetic (as in avoiding 0.75% bupivacaine), or (e) other factors. Most likely, no one factor accounts for this trend, and certainly no single maneuver has reduced complications by itself. No single safety step or test dose, no matter how sensitive or specific, should be regarded as adequate or reliable. Recommendations In view of the information presented, the following observations seem appropriate at this time: - Subdural injection is a rare, but difficult to detect, complication of epidural anesthesia. Only careful attention and a high index of suspicion for the development of the signs of subdural anesthesia will prevent complications. - Subarachnoid injection is a more frequent, but usually more readily recognized, occurrence. It can occur despite negative aspiration. A local anesthetic test dose (usually 3 mL of solution containing a sufficient milligram dose of local anesthetic to produce a rapid onset of subarachnoid block) is indicated before the injection of a larger bolus of epidural anesthetic solution. - Intravascular placement of an epidural needle or catheter remains an issue with epidural anesthesia. Fortunately, the frequency of systemic toxic reactions to local anesthetics seems to have been reduced by the use of multiple safety steps in this situation, including the use of aspiration; incremental injection of local anesthetic; epinephrine-containing test doses in appropriate patients, with appropriate monitoring (heart rate normal, SBP with beta-blockade) and modification of the criteria to fit the clinical scenario (Table 4); alternative test doses (local anesthetic injections [in unmedicated subjects], air) when indicated (advanced age, beta-blockade, pregnancy); and the minimal required concentration and dose of local anesthetic. - No single safety step has been shown to be 100% reliable in preventing systemic toxic reactions. A combination of steps associated with constant vigilance on the part of the anesthesiologist remains the most effective safety measure in performing epidural anesthesia. The authors express their appreciation to Drs. Joseph Neal and Dan Kopacz for critical review of the manuscript, and to the Wood Library Museum for providing the historical texts.