I. Introduction

A. Patient injury associated with regional anesthesia ranges from mere nuisance to life altering. Bruising from axillary brachial plexus block or temporary hoarseness from unintended local anesthetic spread to the recurrent laryngeal nerve may be distressing to the patient, but will resolve fully and quickly. Conversely, clinicians and patients fear permanent nerve injury or death from local anesthetic systemic toxicity (LAST). Life-altering injuries associated with regional anesthesia are extremely rare.

B. Regional anesthesia-related complications are difficult to study because their low frequency makes it difficult to accrue enough patients in randomized controlled trials to achieve sufficient statistical power from which to draw conclusions about incidence, etiology, and risk. Consequently, much of the literature of rare regional anesthesia complications comes from large observational studies or small case reports. Recommendations for avoiding injury are often based on expert opinion derived from analysis of case reports, small case series, or pharmacokinetic inference. Retrospective reviews or voluntary reporting to a central registry suffer from reporting bias (clinicians may choose not to report their serious complications or they may dismiss minor complications as too trivial to merit reporting). Such sources often lack the accuracy, detail, or follow-up necessary to characterize causation, risk, and recovery. Even large prospective studies may fail to ask the right questions or to follow patients long enough to identify late-developing problems. For example, Philips et al.’s (1) classic prospective study of 10,440 patients undergoing lidocaine spinal anesthesia in the 1960s did not detect what came to be recognized in the 1990s as transient neurologic symptoms (TNS) (2).

C. Accurate determination of incidence is difficult for rare complications—the number depends on how the data were derived. For example, in Sweden a young woman undergoing labor analgesia has a 1 per 200,000 risk for epidural hematoma. Yet that same woman at age 70 and undergoing spinal anesthesia for total knee arthroplasty is at 1 per 3,800 risk for the same complication because degenerative spine disease has reduced the cross-sectional area of her spinal canal, thereby leaving little room for both blood and spinal cord (3).

D. Difficulties associated with incidence estimation aside, most contemporary studies report the approximate long-term perioperative nerve injury per regional block as 2-4/10,000 (4); severe LAST as 2.5/10,000 (5); and serious neuraxial infection as 0.2-0.3/10,000 (3).

E. Studies from the Mayo Clinic (6) and elsewhere (7,8) suggest that the decision to perform a regional anesthetic per se does not place patients at higher risk for perioperative nerve injury.

II. Neurologic injury

A. Overview

Central to understanding perioperative nerve injury is appreciating its multifactorial nature. When a needle and drug are used near a nerve that subsequently sustains injury, temporal proximity makes it tempting to blame the complication on the regional anesthetic. Although regional anesthesia can cause neural injury, it is directly responsible for injury in only 10% to 15% of cases (4). The vast majority of perioperative nerve injury is surgery-related,
often from trauma to the nerve, traction, or compression. Patient comorbidity influences risk for perioperative nerve injury, including sex, extremes of body habitus, and medical conditions such as diabetes mellitus, hypertension, and tobacco abuse (8,9).

B. Hemorrhagic complications

1. Spinal hematoma is a potentially devastating complication that can occur spontaneously, but has also been linked to neuraxial block procedures.

2. Specific guidelines for the management of regional anesthetics in patients who are receiving antithrombotic or thrombolytic therapy are complex, updated periodically, and difficult to retain in one’s memory. Therefore, anesthesiologists are referred to the latest version of the American Society of Regional Anesthesia and Pain Medicine’s (ASRA) practice advisory on this topic (10) (www.asra.com) and/or ASRA’s smart phone app (ASRA Coags; iOS or Android).

3. Key concepts related to the prevention of hemorrhagic complications include the following:

a. Multiple anticoagulants increase risk.

b. Traumatic or prolonged needle placement increases risk, but a small amount of blood during placement does not mandate case cancellation.

c. Renal failure increases risk by reducing the clearance of some anticoagulants, particularly the oral anti-Xa inhibitors (rivaroxaban, apixaban, edoxaban) and direct thrombin inhibitors (dabigatran).

d. Bleeding can occur not just with needle or epidural catheter placement, but also with catheter removal.

e. Although aspirin and nonsteroidal anti-inflammatory drugs are considered low risk for regional anesthetic techniques, the same may not apply to large-needle, interventional pain medicine procedures such as intrathecal pumps or spinal cord stimulators.

f. Patients with ankylosing spondylitis or severe spinal stenosis may be at higher risk for neurologic compromise should bleeding occur within the limited cross-sectional area of their spinal canal.

4. All patients with indwelling neuraxial catheters should undergo scheduled neurologic evaluation of lower-extremity sensory and motor function at least every 4 hours. Patients at high risk or heightened concern should be evaluated every 1 to 2 hours.

5. Timely diagnosis of spinal hematoma is paramount because the chances for meaningful recovery diminish rapidly when surgical decompression is delayed more than 8 hours after initial symptom presentation. The severity of neurologic deficit is predictive of ultimate recovery. Back pain and bowel/bladder symptoms may be present, but are not universal. Of particular concern is sensory or motor deficit outside of the block’s expected distribution, for example, lower leg or foot weakness in the presence of a thoracic epidural. Numbness or weakness suspected to arise from local anesthetic action may be addressed by turning the epidural drug infusion down or off, but the patient must be reexamined within an hour. Failure to respond to these measures demands immediate neuroimaging (magnetic resonance imaging [MRI] preferred if immediately available; computerized tomography if not) and referral to a neurosurgeon.

C. Infectious complications

1. Epidural abscess or meningitis can complicate neuraxial anesthetics. Epidural abscess may be indolent initially and present as low-grade fever or mild back pain. Signs and symptoms may then escalate rapidly to include severe back pain, weakness, and/or bowel or bladder dysfunction. Meningitis presents with fever, headache, nuchal rigidity, and/or photophobia.

2. Bacterial colonization at the injection or catheter site is common with peripheral nerve blocks, especially in the axilla or femoral region. True infection is rare—3% or fewer catheters show signs of localized infection, and even fewer develop deep infection that requires drainage or antibiotics.

3. Key to infection prevention is handwashing, jewelry removal, and the use of a face mask during the procedure (11). Blocks should not be performed through infected skin. Data strongly support the superiority of chlorhexidine/alcohol mixtures for skin disinfection.

4. Infection risk is increased when catheters are used for more than 3 to 5 days, in trauma or critical care unit patients, and when the patient develops systemic infection. Regional techniques should not be undertaken, or should be delayed, in patients with untreated sepsis.

5. As with spinal hematoma, correct diagnosis and rapid intervention are paramount to circumvent paralysis or death, which may occur in up to 30% of patients with meningitis.

D. Direct neural trauma

1. Direct injury to the spinal cord or spinal nerves by needles or catheters has been reported as a 5 per million neuraxial anesthetic event (3). The structure that is injured is determined by needle trajectory (Fig. 14.1).

2. Awareness of anatomic factors may lower the risk of direct spinal cord trauma. Risk is increased by the practitioner’s inaccuracy in determining vertebral levels, especially in large patients or those with poor vertebral landmarks. This risk may be mitigated in challenging patients by using ultrasound to determine vertebral level (see Chapter 6 for details). Midline gaps in the mid-to-high thoracic ligamentum flavum occur in up to 20% of patients, thereby negating the expected resistance to epidural needle advancement. The posterior-to-anterior dimension of the epidural space narrows progressively from 4 to 12 mm in the lumbar spine to fewer than 2 mm in the high thoracic spine (12).

3. The spinal cord has no sensory innervation and that to the meninges is inconsistent. Consequently, cases have been reported of needle entry into the spinal cord without warning, even in unsedated patients (12). Injection of substances into the spinal cord usually elicits pain.

4. If direct spinal cord trauma is suspected because of an intense, nonresolving paresthesia or new neurologic deficit, immediate MRI is recommended. If spinal cord injury is confirmed, some experts recommend administering high-dose steroids in consultation with a neurologist or neurosurgeon (4).

5. Needle-related injury of the peripheral nerves is exceptionally rare, in part because nerves tend to move away from approaching needles or catheters. Similar to the spinal cord, paresthesia may or may not herald nerve injury. Injury can occur in the absence of warning signs (9).

E. Neurotoxicity

1. All local anesthetics are neurotoxic and capable of producing permanent neurologic injury. Neurotoxicity increases with local anesthetic concentration and exposure time, that is, reduced drug clearance. Clinically relevant local anesthetic doses and concentrations rarely produce injury in normal individuals or uncompromised neural tissues. (The chapters on regional block techniques recommended doses and concentrations of local anesthetic for specific blocks.) Preservative-free opioids are not neurotoxic.

2. Adjuvants such as epinephrine do not cause nerve injury in isolation, but worsen it by prolonging local anesthetic clearance and thus exposure time. Dexmedetomidine and
clonidine have undergone extensive neurotoxicity studies and are believe to be nontoxic. Dexamethasone is less well studied; some experts advise caution in diabetic patients.

3. Excipients, preservatives, and disinfectants. Commonly used and dosed preservatives or excipients are not neurotoxic (13). There is no evidence that disinfectants such as chlorhexidine/alcohol are neurotoxic when used properly, but they should be allowed to dry before needle placement. Care should be taken to avoid disinfectant splashing onto block trays, needles, or drugs.

4. Extreme care should also be taken to avoid injection of unintended substances during single-injection techniques or into continuous delivery systems. Precautions include proper labelling of syringes and drugs, reading and re-reading labels before injection, and avoidance of superfluous stopcocks or port access on catheters that subserve a neural structure.

a. The precise etiology of regional anesthetic-associated peripheral nerve injury is unknown. Classic theory suggests a dual mechanism by which a needle or catheter disrupts the perineurium, thereby exposing denuded axons to the neurotoxic effects of local anesthetic (which is nontoxic to intact axons) (Fig. 14.2).

b. An additional postulated etiology of peripheral nerve injury involves localized or diffuse inflammation (see Postsurgical Inflammatory Neuropathies)

c. Because needle or catheter injury may expose the axons to neurotoxic agents, most experts recommend avoiding needle-to-nerve contact during peripheral nerve block placement. Nevertheless, ultrasound studies of patients in whom initial nerve localization was accomplished using peripheral nerve stimulation (PNS) confirm unintended intraneural needle placement more often than appreciated previously. The fact that these patients rarely suffer injury is likely a function of the high proportion of nonneural connective tissue present within a nerve’s cross-sectional microanatomy, resistance of the perineurium to needle penetration, and acknowledgment that direct needle injury may not play as significant a role in nerve injury as once thought.

d. Although ultrasound guidance (UG) helps avoid unintended subepineurial needle placement, there is no evidence that ultrasound use per se has reduced the incidence of anesthesia-related nerve injury from that previously reported with PNS alone (14). Some experts advocate localizing peripheral nerves with both UG and PNS. This opinion is based on extrapolated evidence from human supraclavicular and popliteal sciatic block studies that demonstrate a motor response elicited at 0.2 mA or less is consistent with subepineurial needle placement, whereas a response at 0.5 mA or greater is most often indicative of extraneural needle placement (9). Nevertheless, there is no evidence that combining UG with PNS reduces nerve injury, but it does increase procedural time without concomitant increasing block success (Fig. 14.3).

e. Patient report of paresthesia or pain associated with needle-to-nerve contact or unintended intraneural injection is inconsistent, but when it happens serves as yet another safety monitor. For this reason, the use of regional anesthesia in anesthetized or deeply sedated adults is discouraged (4). However, regional anesthesia is performed routinely in anesthetized children.

(1) Postoperative neurologic symptoms after peripheral nerve block are reported in up to 20% of patients the day after surgery, diminish to 3% at 1 month, and to 2-4/10,000 at 9 to 12 months.

(2) Most sensory symptoms will resolve within days to weeks and can be managed expectantly.

(3) More worrisome signs and symptoms include complete nerve deficit at the end of surgery, motor signs, recrudescence of block after initial resolution, and/or severe pain. Any of these presentations should be treated as an emergency because expanding hematoma, constricting bandages or casts, and so forth can often be rectified with full symptom resolution.

(4) When the injury fails to resolve or has worrisome components, early referral to a neurologist is warranted. Although electrophysiologic studies do not change for 2 to 3 weeks after an injury, consideration may be given to early, bilateral studies when the injury is severe.

(5) Neuropathic pain should be treated with opioids and/or gabapentinoids, plus referral to a neurologist or chronic pain physician.

a. The cauda equina may be particularly susceptible to neurotoxic injury because intrathecal nerve roots have no protective covering and a long transit course, thereby increasing their surface area (Fig. 14.4). Cauda equina syndrome (CES) from presumed local anesthetic neurotoxic injury is rare and poorly understood. One study observed CES in 0.2/10,000 neuraxial anesthetics. In this report, 28% of CES cases were associated with spinal stenosis, whereas the remaining cases were unremarkable in terms of local anesthetic dose, concentration, adjuvant use, and subsequent neuroimaging (3).

b. CES has been linked to supernormal doses of local anesthetic and/or local anesthetic maldistribution within the lumbosacral intrathecal sac, as manifested by dense sacral but no lumbar-thoracic anesthesia.

c. CES is also postulated to be the result of inflammation brought on by the anesthetic or surgical procedure, the drugs used, patient characteristics, or yet unknown factors.

d. Because of CES association with supernormal doses, it is recommended that failed spinal anesthetics not be redosed, or if they are, that the total local anesthetic dose does not exceed the recommended maximum for a single injection.

F. Spinal cord ischemia and infarction

1. Ischemic injury or infarction of the spinal cord is a decidedly rare and poorly understood event. Although ischemia and infarction can affect any portion of the spinal cord, the most common variety is anterior spinal artery syndrome (ASAS). Most ASAS presents spontaneously in patients at risk for vascular disease (hypertension, tobacco abuse, advanced age). When ASAS or other forms of spinal ischemia present in the setting of a neuraxial block, the latter is often invoked as the causative event.

2. ASAS presents classically as sudden or rapidly progressive motor and sensory deficit (pain and temperature) with preserved posterior column functions (proprioception, vibration, fine touch). Varied presentations can occur, such as slower, progressive onset rather than sudden lower-extremity muscle flaccidity.

3. Key points related to spinal cord ischemia include the following:

a. Direct needle injury to the anterior spinal artery cannot occur without the needle traversing the spinal cord. However, needles and catheters placed lateral to the vertebrae may injure spinal arteries or radicular arteries that supply the spinal cord (Fig. 14.5).

b. Because the lower thoracic and lumbosacral spinal cord is supplied by the single arterius radicularis magnus, deficits involving the lower spinal cord and conus medullaris are most common.

c. Despite being invoked as the agent that causes ischemia, epinephrine in clinically used concentrations does not adversely affect spinal cord blood flow (SCBF) (15).