Although the most common adverse reactions to local anesthetics are vasovagal events associated with injection,113 the following sections focus on their local and systemic toxicity.
Regional Side Effects and Tissue Toxicity
At a sufficient concentration, all local anesthetics are directly cytotoxic to nerve cells. However, in clinically relevant doses, they rarely produce localized nerve damage.58,83 Significant direct neurotoxicity may result from intrathecal injection or infusion of local anesthetics for spinal anesthesia. In this setting, lidocaine has an increased risk for both persistent lumbosacral neuropathy and a syndrome of painful but self-limited postanesthesia buttock and leg pain or dysesthesia referred to as transient neurologic symptoms.53 Nerve damage often is attributed to use of excessively concentrated solutions or inappropriate formulations. Several reports of cauda equina syndrome are associated with use of hyperbaric 5% lidocaine solutions for spinal anesthesia. Hyperbaric solutions are denser than cerebrospinal fluid. This neurotoxicity appears to be a phenomenon that occurs when the anesthetic is injected through narrow-bore needles or through continuous spinal catheters. This process may result in very high local concentrations of the anesthetic that might pool around the sacral roots because of inadequate mixing.97 The mechanism of this neurotoxicity is unknown but is believed to be independent of sodium channel blockade.53 Because an equally effective block can be achieved with injection of larger volumes of lower concentration, 5% lidocaine should be avoided and bupivacaine used instead. There is a significant (up to 10-fold) increase in the development of new neurologic dysfunction after receiving a neuraxial block in patients with preexisting peripheral neuropathy a fact emphasizing informed consent.46
Similar severe neurotoxic reactions occur after massive subarachnoid injection of chloroprocaine during attempted epidural anesthesia.94 The neurotoxicity initially appeared to be associated with use of the antioxidant sodium bisulfite and the low pH of the commercial solution rather than use of the anesthetic itself.115 Although chloroprocaine has been reformulated without bisulfite, new animal data suggest that it is the anesthetic itself that may be responsible for the neurotoxicity.111 Skeletal muscle changes are observed after intramuscular injection of local anesthetics, especially the more potent, longer acting xenobiotics. The effect is reversible, and muscle regeneration is complete within 2 weeks after injection of local anesthetics.8
Although rare, transient or prolonged postoperative neuropathy after peripheral nerve or plexus block is well recognized. Likely mechanisms include direct injury of the nerve related to intraneuronal injection and local anesthetic neurotoxicity. The frequency of peripheral neuropathies reported after peripheral nerve blockade varies from 0% to more than 5%.12 The use of ultrasound guidance to direct needle position may reduce this complication because of direct visualization of the peripheral nerve, avoiding traumatic injury and allowing for injection of less local anesthetic to produce adequate nerve block. Comparative studies have shown conflicting results and were likely underpowered. Recently, analysis of 12,868 ultrasound-guided nerve blocks for peripheral regional anesthesia found that the incidences of postoperative neurologic symptoms lasting longer than 5 days and 6 months were 0.18% and 0.008%, respectively.105 Visualization of the target site via ultrasound guidance may well be shown to also decrease local anesthetic systemic toxicity by averting inadvertent intravascular injection.
Systemic Side Effects and Toxicity
Allergic reactions to local anesthetics are extremely rare. Fewer than 1% of all adverse drug reactions caused by local anesthetics are immunoglobulin (Ig) E mediated.40 In one study designed to determine the prevalence of true local anesthetic allergy in patients referred to an allergy clinic for suspected hypersensitivity, skin prick and intradermal testing results were negative for all 236 subjects tested.9 As noted, the amino esters are responsible for the majority of true allergic reactions. When hydrolyzed, the amino ester local anesthetics produce PABA, a known allergen (Chap. 55). Cross-sensitivity to other amino ester anesthetics is common. Some multidose commercial preparations of amino amides may contain the preservative methylparabens (Chap. 55), which is chemically related to PABA and is the most likely cause of the much rarer allergic reactions attributed to amino amides. Preservative-free amino amides, including lidocaine, can be used safely in patients who have reactions to drug preparations containing methylparabens unless the patient is specifically sensitive to lidocaine. Again, if the patient with a history of allergic reaction to a particular drug requires a local anesthetic, a paraben preservative–free drug from the opposite class can be chosen because there is no cross-reactivity between the amides and esters.
Methemoglobinemia is a frequent adverse effect of topical and oropharyngeal benzocaine and is occasionally reported with lidocaine, tetracaine, or prilocaine use. The diagnosis can be established by direct measurement of methemoglobin with a cooximeter. Most reports of methemoglobinemia associated with local anesthetics are the result of an excessive dose or a break in the normal mucosal barrier for topical anesthetics (Chap. 127).
Benzocaine is metabolized to aniline and then further metabolized to phenylhydroxylamine and nitrobenzene, which are both potent oxidizing agents (Chap. 127). Although reports describe methemoglobinemia resulting from standard doses of benzocaine topical oropharyngeal spray given for laryngoscopy or gastrointestinal upper endoscopy,30,79 affected patients commonly have abnormal mucosal integrity as occurs with thrush or mucositis. Prilocaine is an amino ester local anesthetic primarily used in obstetric anesthesia because of its rapid onset of action and low systemic toxicity in both the mother and fetus. Use of large doses of prilocaine may lead to the development of methemoglobinemia.48,65 Prilocaine is an aniline derivative that, when metabolized in the liver, produces ortho-toluidine, another oxidizing agent.48 A direct relationship exists between the amount of epidural prilocaine administered and the incidence of methemoglobinemia. A dose greater than approximately 8 mg/kg is generally necessary to produce symptoms, which may not become apparent until several hours after epidural administration of the drug. EMLA cream, often used in the outpatient setting for minor dermal procedures, may result in significant methemoglobinemia, which has been reported in children and rarely in adults.42 Standard doses of EMLA cream used for circumcision in term neonates are associated with minimal production of methemoglobin, but risks may be increased in neonates with metabolic disorders.109 When clinically indicated, affected patients with symptomatic methemoglobinemia should be treated with IV methylene blue (Chap. 127 and Antidotes in Depth: A42).
Systemic toxicity for all local anesthetics correlates with serum concentrations. Factors that determine the concentration include: dose; rate of administration; site of injection (absorption occurs more rapidly and completely from vascular areas, such with neck blocks and intercostal blocks); the presence or absence of a vasoconstrictor; and the degree of tissue–protein binding, fat solubility, and pKa of the local anesthetic.78 The brain and heart are the primary target organs for systemic toxicity because of their rich perfusion, moderate tissue–blood partition coefficients, lack of diffusion limitations, and presence of cells that rely on voltage-gated sodium channels to produce an action potential.
Recommendations for maximal local anesthetic doses designed to minimize the risk for systemic toxic reactions are published.107 These maximal recommended doses aim to prevent infiltration of excessive drug. However, because most episodes of systemic toxicity from local anesthetics, with the exception of methemoglobinemia from topical drug, occur secondary to unintentional intravascular injection rather than from overdosage, limiting the maximal dose will not prevent most toxic systemic reactions.102
Toxicity is also related to the metabolism for a given local anesthetic. The rapidity of elimination from the plasma influences the total dose delivered to the CNS or heart. The amino esters are rapidly hydrolyzed in the plasma and eliminated, explaining their relatively low potential for systemic toxicity. The amino amides have a much greater potential for producing systemic toxicity because termination of the therapeutic effect of these drugs is achieved through redistribution and slower metabolic inactivation.36 Another factor that creates difficulty in specifying the minimal toxic plasma concentration of lidocaine results from the fact that its N-dealkylated metabolites are pharmacologically active. Although these factors make it difficult to establish safe doses of local anesthetics, Table 67–2 summarizes the estimates of minimal toxic IV doses of various local anesthetics.
Table 67–2.Toxic Intravenous (IV) Doses of Local Anesthetics ||Download (.pdf) Table 67–2. Toxic Intravenous (IV) Doses of Local Anesthetics
Central Nervous System Toxicity
Systemic toxicity in humans usually presents with CNS abnormalities. IV infusion studies in volunteers demonstrate an inverse relationship between anesthetic potency and dose required to induce signs of CNS toxicity.104 A similar relationship exists between the convulsive concentration and the relative anesthetic potency. In humans, seizures are reported at serum concentrations of approximately 2 to 4 μg/mL for bupivacaine and etidocaine. Concentrations in excess of 10 μg/mL are usually required for production of seizures when less potent drugs such as lidocaine are administered. Despite the strong relationship between local anesthetic potency and CNS toxicity, several other factors influence the CNS effects, including the rate of injection, drug interactions, and acid–base status.25
The rapidity with which a particular serum concentration is achieved influences the toxicity of the anesthetic. Volunteers could tolerate an average dose of 236 mg of etidocaine and a serum concentration of 3 μg/mL before onset of CNS symptoms when the anesthetic was infused at a rate of 10 mg/min. However, when the infusion rate was increased to 20 mg/min, the same individuals could tolerate only an average of 161 mg of the drug, which produced a serum concentration of approximately 2 μg/mL.103
Centrally acting local anesthetics can modify the clinical presentation of a systemic toxic reaction. In general, CNS-depressant drugs minimize the signs and symptoms of CNS excitation and increase the threshold for local anesthetic–induced seizures. Flumazenil increases the sensitivity of the CNS to the amino amide anesthetics.14
Both metabolic and respiratory acidoses increase local anesthetic–induced CNS toxicity. Acidemia decreases plasma protein binding, increasing the amount of free drug available for CNS diffusion despite promoting the charged form of the amine group. The convulsive threshold of various local anesthetics is inversely related to arterial PCO2.27,33,34 Hypercarbia may lower the seizure threshold by several mechanisms: (1) increased cerebral blood flow, which increases drug delivery to the CNS; (2) increased conversion of the drug base to the active cation in the presence of decreased intracellular pH; and (3) decreased plasma protein binding, which increases the amount of free drug available for diffusion into the brain.16,27,33,34
A gradually increasing serum lidocaine concentration usually produces a stereotypical pattern of symptoms and signs (Fig. 67–3). In an awake patient, the initial effects include tinnitus, lightheadedness, circumoral numbness, disorientation, confusion, auditory and visual disturbances, and lethargy. Subjective side effects occur at serum concentrations between 3 and 6 μg/mL. Significant psychological effects of local anesthetics are also reported. Near-death experiences and delusions of actual death are described as specific symptoms of local anesthetic toxicity.68 Thus, the appearance of psychological symptoms during administration of local anesthetics should not be disregarded as unrelated nervous reactions or effects of sedatives given as premedication but rather as a possible early sign of CNS toxicity.
Relationship of signs and symptoms of toxicity to serum lidocaine concentrations.
Clinical signs, usually excitatory, then develop, and include shivering, tremors, and ultimately generalized tonic–clonic seizures. Objective CNS toxicity usually is evident at lidocaine concentrations between 5 and 9 μg/mL. Seizures may occur at concentrations above 10 μg/mL, with higher concentrations producing coma, apnea, and cardiovascular collapse. The excitatory phase has a wide range of intensity and duration, depending on the chemical properties of the local anesthetic. With the highly lipophilic, highly protein-bound drugs, the excitement phase is brief and mild. Toxicity from a large IV bolus of bupivacaine may present without any CNS excitement, with bradycardia, cyanosis, and coma as the first signs.99 Rapid intravascular injection of lidocaine may produce a brief excitatory phase followed by generalized CNS depression with respiratory arrest. Seizures may follow even a small dose injected into the vertebral or carotid artery (as may occur during stellate ganglion block).57 A relative overdose produces a slower onset of effects (usually within 5–15 minutes of drug injection), with irritability progressing to seizures.
The mechanism of the initial CNS excitation involves a selective block of cerebral cortical inhibitory pathways in the amygdala.110,114 The resulting increase in unopposed excitatory activity leads to seizures. As the concentration increases further, both inhibitory and excitatory neurons are blocked, and generalized CNS depression ensues.
Treatment of Local Anesthetic Central Nervous System Toxicity
At the first sign of possible CNS toxicity, administration of the drug must be discontinued. One hundred percent oxygen should be supplied immediately, and ventilation should be supported if necessary. Patients with minor symptoms usually do not require treatment, provided adequate respiratory and cardiovascular functions are maintained. The patient must be followed closely so that progression to more severe effects can be detected.
Although most seizures caused by local anesthetics are self-limited, they should be treated quickly because the hypoxia and acidemia produced by prolonged seizures may increase both CNS and cardiovascular toxicity.80,82 Intubation is not mandatory, and the decision to intubate must be individualized. Maintaining adequate ventilation is of proven value, but modest hyperventilation, in theory, might decrease CNS toxicity. By decreasing CNS extraction of drug, lowering extracellular potassium, and hyperpolarizing the neuronal cell membrane, normalizing (lowering) PCO2 may decrease the affinity or accelerate separation of the local anesthetic from the sodium channel. Ultra-short-acting barbiturates and benzodiazepines have been used for treatment of local anesthetic–induced seizures, but either of these medication groups can also exacerbate circulatory and respiratory depression.25,75 Propofol 1 mg/kg IV was as effective as thiopental 2 mg/kg IV in stopping bupivacaine induced seizures in rats and has been used successfully in a patient with uncontrolled muscle twitching secondary to local anesthetic toxicity.10,44 However, propofol may cause significant bradydysrhythmias and even asystole, especially when used with other xenobiotics that cause bradycardia. Whether propofol interacts with local anesthetics to enhance their bradydysrhythmic effects is not known, and it is not possible to generally recommend propofol over benzodiazepines for treatment of local anesthetic CNS toxicity. Neuromuscular blockers are proposed as adjunctive treatment for local anesthetic induced seizures. They block muscular activity, decreasing oxygen demand and lactic acid production. However, neuromuscular blockers should never be used to treat seizures per se because they have no anticonvulsant effect and can make clinical diagnosis of ongoing seizures problematic by abolishing muscle contractions. To avoid this potentially lethal complication, chemical paralysis should be used only to facilitate endotracheal intubation if needed, unless continuous electroencephalography is also used. If used, short-acting neuromuscular blockers are desirable, facilitating subsequent repeated neurologic assessments. Succinylcholine may not be ideal because of its significant side effects, including hyperkalemia and dysrhythmias. The use of nondepolarizing neuromuscular blockers with less potential for cardiac side effects, such as rocuronium, should be considered (Chap. 69).
When severe systemic toxicity occurs, the cardiovascular system must be monitored closely because cardiovascular depression may go unnoticed while seizures are being treated. Because local anesthetic–induced myocardial depression may occur even with preserved blood pressure, it is important to be aware of early signs of cardiac toxicity, including electrocardiographic (ECG) changes.
If toxicity results from ingestion of liquid medications, as most are, activated charcoal is generally indicated, but benefits are unproven. If the patient presents immediately after ingestion, gastric lavage with a nasogastric tube may be considered. Induction of emesis is contraindicated even after oral administration because of the risk of seizures and aspiration. Contaminated mucous membranes should be washed off. Hemodialysis is not of proven utility and may be impractical, as is hemoperfusion.
Cardiovascular side effects are the most feared manifestations of local anesthetic toxicity. Shock and cardiovascular collapse may be related to effects on vascular tone, inotropy, and dysrhythmias related to indirect CNS and direct cardiac and vascular effects of the local anesthetic. Animal studies and clinical observations clearly demonstrate that for most local anesthetics, CNS toxicity develops at significantly lower serum concentrations (exception: bupivacaine) than those needed to produce cardiac toxicity, that is, they have a high cardiovascular:CNS toxicity ratio.57,81,82,99 When cardiac toxicity occurs, management may be exceedingly difficult. Some of the discrepancy between the incidence of CNS and cardiac toxicity may result from a detection bias. Not only can the treating physicians fail to recognize cardiac effects because of preoccupation with CNS manifestations of toxicity, but early cardiac toxicity may be quite subtle. An experimental study attempting to identify early warning signs of bupivacaine-induced cardiac toxicity in pigs evaluated bupivacaine-induced changes in cardiac output, heart rate, blood pressure, and ECG.88 A 40% reduction in cardiac output was not associated with significant change in heart rate or blood pressure, the latter secondary to a direct vasoconstrictive effect of bupivacaine at the concentrations produced.18
Changes in systemic vascular tone induced by local anesthetics may be mediated by direct effect on vascular smooth muscle or indirectly via effects on spinal cord sympathetic outflow. Predictably, sympathetic blockade after spinal anesthesia or epidural anesthesia above the T5 dermatome results in peripheral venodilation and arterial dilation. Shock may result when high doses of anesthetic are used in hypovolemic patients. Local anesthetics have a biphasic effect on peripheral vascular smooth muscle. Whereas lower doses produce direct vasoconstriction, higher doses are associated with severe cardiovascular toxicity and cause vasodilation, contributing to cardiovascular collapse.
All local anesthetics directly produce a dose dependent decrease in cardiac contractility, with the effects roughly proportional to their peripheral anesthetic effect. Although the classic anesthetic action of sodium channel blockade in heart muscle accounts in large part for the negative inotropy by affecting excitation–contraction coupling, it does not explain the entire difference in myocardial depression produced by different anesthetics.28 Poorly understood effects on calcium handling or effects of the intracellular drug directly on contractile proteins or mitochondrial function may be operable.28
Blockade of the fast sodium channels of cardiac myocytes decreases maximum upstroke velocity (Vmax) of the action potential (Chaps. 16 and 17 and Fig. 64–1). This effect slows impulse conduction in the sinoatrial and atrioventricular (AV) nodes, the His-Purkinje system, and atrial and ventricular muscle.21 These changes are reflected on ECG by increases in PR interval and QRS duration. At progressively higher anesthetic concentrations, hypotension, sinus arrest with junctional rhythm, and eventually cardiac arrest occur.4 Asystole has been described in patients who received unintentional IV bolus injections of 800 to 1000 mg of lidocaine.4,35 Cardiovascular toxicity of local anesthetics usually occurs after a sudden increase in serum concentration, as in unintentional intravascular injection. Cardiovascular toxicity is rare in other circumstances because high serum concentrations are necessary to produce this effect and because CNS toxicity precedes cardiovascular events, providing a warning. Cardiac toxicity usually is not observed with lidocaine use in humans until the serum lidocaine concentration greatly exceeds 10 μg/mL unless the patient is also receiving xenobiotics that depress sinus and AV nodal conduction such as calcium channel blockers, β-adrenergic antagonists, or cardioactive steroids.
Bupivacaine is significantly more cardiotoxic than most other local anesthetics commonly used. Inadvertent intravascular injection produces near simultaneous signs of CNS and cardiovascular toxicity.
Animal studies have compared the dose or serum concentrations of local anesthetics required to produce irreversible circulatory collapse with those necessary to produce seizures.26,81,82 This cardiovascular collapse:CNS toxicity (CC:CNS) ratio for lidocaine is approximately 7; therefore, CNS toxicity should become evident well before potentially cardiotoxic concentrations are reached. In contrast, the CC:CNS ratio for bupivacaine is 3:7. Bupivacaine produces myocardial depression out of proportion to its anesthetic potency and, more important, may cause refractory ventricular dysrhythmias.101 Enhanced cardiovascular toxicity may relate to enhanced CNS effects at cardiovascular centers,112 direct effects on myocyte metabolism, and important differences related to sodium channel blockade. Although lidocaine and bupivacaine both block sodium channels in the open or inactivated states, lidocaine quickly dissociates from the channel at diastolic potentials, allowing rapid recovery from block during diastole (fast on–fast off ). Therefore, sodium channel blockade with lidocaine is much more pronounced at rapid heart rates (accounting for the antidysrhythmic effects for ventricular tachycardia).67 On the other hand, at high concentrations, bupivacaine rapidly binds to and slowly dissociates from sodium channels (fast on–slow off), with significant block accumulating at all physiologic heart rates.21 Accordingly, at heart rates of 60 to 150 beats/min, approximately 70 times more lidocaine is needed than bupivacaine to produce an equal effect on Vmax of the action potential. Enhanced conduction block in Purkinje fibers and ventricular muscle cells sets up a reentrant circuit responsible for the ventricular tachydysrhythmias induced by bupivacaine.74
Bupivacaine, a potent and long-acting amide anesthetic, not only has the highest potential for cardiovascular toxicity, but these effects are quite refractory to conventional therapy. Bupivacaine has an asymmetrically substituted carbon, and the kinetics of sodium channel binding are stereospecific.60 The S (levo)-enantiomer levobupivacaine is significantly less cardiotoxic than the R (dextro)-enantiomer despite having similar anesthetic properties.6,72 Consequently, bupivacaine, the racemic mixture of both enantiomers, is more cardiotoxic than levobupivacaine, which contains only the levo-enantiomer.41 The stereospecific effect on sodium channels seems to differ between the heart and the peripheral nerves because the local anesthetic potency of levobupivacaine is the same as, or perhaps even greater than, that of bupivacaine.31,85 Ropivacaine is a pure enantiomer and is less cardiotoxic than bupivacaine, but it is also slightly less potent as an anesthetic.91,92
Effects other than sodium channel blockade may contribute to cardiotoxicity. Lipophilic local anesthetics such as bupivacaine may directly impair mitochondrial energy transduction via two mechanisms: (1) uncoupling of oxygen consumption and adenosine triphosphate (ATP) synthesis and (2) inhibition of complex I in the respiratory chain.101 This effect is related to the lipophilic properties of the drug rather than to stereospecific effects on ion channels. Lidocaine has no effect on mitochondrial respiration, and ropivacaine has less effect than bupivacaine.121 There is no difference between the two bupivacaine enantiomers. These effects occur with higher concentrations of the local anesthetic, as occur after unintentional intravascular injection.
Low dose bupivacaine-induced cardiotoxic effects are described in humans under certain circumstances and at concentrations that are not associated with seizure activity in pigs.55,120 Severe cardiac toxicity is described after injection of a small subcutaneous dose of bupivacaine in a patient with secondary carnitine deficiency.120 Myocytes are highly dependent on oxidation of free fatty acids for energy. Interference with this mechanism via bupivacaine-induced inhibition of carnitine-acylcarnitine translocase has been proposed to contribute to the cardiotoxicity of lipophilic local anesthetics120 (Chap. 48, Fig. 48–2, and Antidotes in Depth: A20). Bupivacaine may produce dysrhythmias by blocking GABAergic neurons that tonically inhibit the autonomic nervous system.45 In addition to its other effects on the heart, bupivacaine may induce a marked decrease in cardiac contractility by altering Ca2+ release from sarcoplasmic reticulum.66
In a large series of patients receiving bupivacaine, systemic toxicity occurred in only 15 of 11,080 nerve blocks.77 Of these patients, 80% convulsed; the other 20% had milder symptoms. A series of cases was described in which bupivacaine use, particularly at 0.75% concentration, was associated with severe cardiovascular depression, ventricular dysrhythmias, and even death. Pregnant women were disproportionately affected. Some of these patients required prolonged resuscitation, and restoration of adequate spontaneous circulation proved exceedingly difficult.95 In 1983, 49 incidents of cardiac arrest or ventricular tachycardia that occurred over a 10-year period were presented to the US FDA Anesthetic and Life Support Advisory Committee. Among these cases, 0.75% bupivacaine was used in 27 obstetric patients with 10 deaths, and 0.5% bupivacaine was used in 8 obstetric patients with 6 deaths. Among the 14 nonobstetric patients, 5 died. The overall mortality rate was 21 of 49 (43%). Partly as a result of these reports, in 1984, the FDA withdrew approval of bupivacaine 0.75% for use as obstetric anesthesia.95
Acid–base and electrolyte status influence the cardiac toxicity of a given drug because all depressant properties are potentiated by acidosis, hypoxia, or hypercarbia.13 Table 67–3 outlines the spectrum of acute local anesthetic reactions.
Table 67–3.Types of Local Anesthetic Reactions ||Download (.pdf) Table 67–3. Types of Local Anesthetic Reactions
|Cause ||Major Clinical Features |
|Local anesthetic toxicity (intravascular injection) ||Immediate seizure or dysrhythmias |
|Reaction to catecholamine ||Tachycardia, hypertension, headache |
|Vasovagal reaction ||Bradycardia, rapid onset and recovery, hypotension, pallor |
|Allergic reaction ||Anaphylaxis |
|High spinal or epidural block ||Bradycardia, hypotension, respiratory distress, respiratory arrest |