Chapter 45: Spinal, Epidural, & Caudal Blocks

Spinal, caudal, and epidural blocks were first used for surgical procedures at the turn of the twentieth century. These central blocks were widely used worldwide until reports of permanent neurological injury appeared, most prominently in the United Kingdom. However, a large-scale epidemiological study conducted in the 1950s proved that complications were rare when these blocks were performed skillfully, with attention to asepsis, and when newer, safer local anesthetics were used. Today, neuraxial blocks are routinely employed for labor analgesia, cesarian delivery, orthopedic surgery, perioperative analgesia, and chronic pain management. However, they are still associated with various complications, and much literature has examined the incidence of complications following neuraxial blocks associated with different disease states.

Neuraxial anesthesia greatly expands the anesthesiologists’ armamentarium, in many cases providing alternatives to general anesthesia. Neuraxial anesthesia may be used simultaneously with general anesthesia or afterward for postoperative analgesia. Neuraxial blocks can be performed as a single injection or with a catheter to allow intermittent boluses or continuous infusions.

Adverse reactions and complications associated with regional anesthesia range from self-limited back soreness to debilitating permanent neurological deficits and even death. The practitioner must therefore be thoroughly familiar with the anatomy involved and the pharmacology and toxic dosages of the agents employed. The practitioner must diligently employ sterile techniques and quickly address physiological derangements arising from neuraxial techniques.

Some studies suggest that postoperative morbidity may be reduced when neuraxial blockade is used either alone or in combination with general anesthesia. Some less convincing studies suggest that neuraxial blocks are associated with reduced perioperative mortality. Neuraxial blocks may reduce the incidence of venous thrombosis and pulmonary embolism, cardiac complications in high-risk patients, bleeding and transfusion requirements, vascular graft occlusion, and pneumonia and respiratory depression following upper abdominal or thoracic surgery in patients with chronic lung disease. Neuraxial blocks may also allow earlier return of gastrointestinal function following surgery. Proposed mechanisms (in addition to avoidance of larger doses of anesthetics and opioids) include reducing the hypercoagulable state associated with surgery, increasing tissue blood flow, improving oxygenation from decreased splinting, enhancing peristalsis, and suppressing the neuroendocrine stress response to surgery. Reduction of parenteral opioid administration may decrease the incidence of atelectasis, hypoventilation, and aspiration pneumonia and reduce the duration of ileus. Postoperative epidural analgesia may also significantly reduce both the need for mechanical ventilation and the time until extubation after major abdominal or thoracic surgery.

The Sick Elderly Patient

Anesthesiologists are all too familiar with situations in which a consultant “clears” a sick elderly patient with significant cardiac disease for surgery “under spinal anesthesia.” A spinal anesthetic with little or no intravenous sedation may reduce the likelihood of postoperative delirium or cognitive dysfunction, which is sometimes seen in the elderly. Unfortunately, many patients will require some sedation during the procedure, either for comfort or to facilitate cooperation. Is spinal anesthesia always safer in a patient with severe coronary artery disease or with a decreased ejection fraction? Ideally, an anesthetic technique should produce neither hypotension (which decreases myocardial perfusion pressure) nor hypertension and tachycardia (which increase myocardial oxygen consumption). Spinal anesthesia may produce both hypotension and bradycardia. Administration of large intravenous volumes may lead to fluid overload in the elderly patient with diastolic dysfunction, especially after the sympathetic block resolves postoperatively. General anesthesia, on the other hand, also poses potential problems for patients with cardiac compromise. Most general anesthetics are cardiac depressants, and many cause vasodilation. Deep anesthesia can readily cause hypotension, whereas light anesthesia relative to the level of stimulation causes hypertension and tachycardia. Research is ongoing to discern if neuraxial techniques offer survival and other benefits to patients compared with general anesthetics for major operations such as open reduction and internal fixation of femoral neck fractures. To date, the results are conflicting.

The Obstetric Patient

Currently, epidural anesthesia is widely used for analgesia in women in labor and during vaginal delivery. Cesarean delivery is most commonly performed under spinal or epidural anesthesia. Both blocks allow a mother to remain awake and experience the birth of her child. Large population studies in Great Britain and the United States have shown that regional anesthesia for cesarean delivery is associated with less maternal morbidity and mortality than is general anesthesia. This may be largely due to a reduction in the incidence of pulmonary aspiration and failed intubation when neuraxial anesthesia is employed. Fortunately, the increased availability of video laryngoscopes may also reduce the incidence of adverse outcomes related to airway difficulties associated with general anesthesia for cesarean delivery.

THE VERTEBRAL COLUMN

The spine is composed of the vertebral bones and intervertebral disks (Figure 45–1). There are 7 cervical (C), 12 thoracic (T), and 5 lumbar (L) vertebrae (Figure 45–2). The sacrum is a fusion of 5 sacral (S) vertebrae, and there are small rudimentary coccygeal vertebrae. The spine as a whole provides structural support for the body and protection for the spinal cord and nerves and allows a degree of mobility in several spatial planes. At each vertebral level, paired spinal nerves exit the central nervous system (Figure 45–2).

Vertebrae differ in shape and size at the various levels. The first cervical vertebra, the atlas, lacks a body and has unique articulations with the base of the skull and with the second vertebra. The second vertebra, called the axis, consequently has atypical articulating surfaces. All 12 thoracic vertebrae articulate with their corresponding rib. Lumbar vertebrae have a large anterior cylindrical vertebral body. A hollow ring is defined anteriorly by the vertebral body, laterally by the pedicles and transverse processes, and posteriorly by the lamina and spinous processes (Figure 45–1B and C). The laminae extend between the transverse processes and the spinous processes, and the pedicle extends between the vertebral body and the transverse processes. When stacked vertically, the hollow rings become the spinal canal in which the spinal cord and its coverings sit. The individual vertebral bodies are connected by the intervertebral disks. There are four small synovial joints at each vertebra, two articulating with the vertebra above it and two with the vertebra below. These are the facet joints, which are adjacent to the transverse processes (Figure 45–1C). The pedicles are notched superiorly and inferiorly, these notches forming the intervertebral foramina from which the spinal nerves exit. Sacral vertebrae normally fuse into one large bone, the sacrum, but each one retains discrete anterior and posterior intervertebral foramina. The laminae of S5 and all or part of S4 normally do not fuse, leaving a caudal opening to the spinal canal, the sacral hiatus (Figure 45–3).

The spinal column normally forms a double C, being convex anteriorly in the cervical and lumbar regions (Figure 45–2). Ligamentous elements provide structural support, and, together with supporting muscles, help to maintain the unique shape. Ventrally, the vertebral bodies and intervertebral disks are connected and supported by the anterior and posterior longitudinal ligaments (Figure 45–1A). Dorsally, the ligamentum flavum, interspinous ligament, and supraspinous ligament provide additional stability. Using the midline approach, a needle passes through these three dorsal ligaments and through an oval space between the bony lamina and spinous processes of adjacent vertebra (Figure 45–4).

THE SPINAL CORD

The spinal canal contains the spinal cord with its coverings (the meninges), fatty tissue, and a venous plexus (Figure 45–5). The meninges are composed of three layers: the pia mater, the arachnoid mater, and the dura mater; all are contiguous with their cranial counterparts (Figure 45–6). The pia mater is adherent to the spinal cord, whereas the arachnoid mater is usually adherent to the thicker and denser dura mater. Cerebrospinal fluid (CSF) is contained between the pia and arachnoid maters in the subarachnoid space. The spinal subdural space is generally a poorly demarcated, potential space that exists between the dura and arachnoid membranes. The epidural space is a better defined potential space within the spinal canal that is bounded by the dura and the ligamentum flavum (Figures 45–1 and 45–5).

The spinal cord normally extends from the foramen magnum to the level of L1 in adults (Figure 45–7). In children, the spinal cord ends at L3 and moves up with age. The anterior and posterior nerve roots at each spinal level join one another and exit the intervertebral foramina, forming spinal nerves from C1 to S5 (Figure 45–2). At the cervical level, the nerves arise above their respective vertebrae, but starting at T1, exit below their vertebrae. As a result, there are eight cervical nerve roots, but only seven cervical vertebrae. The cervical and upper thoracic nerve roots emerge from the spinal cord and exit the vertebral foramina nearly at the same level (Figure 45–2). But, because the spinal cord normally ends at L1, lower nerve roots course some distance before exiting the intervertebral foramina. These lower spinal nerves form the cauda equina (“horse’s tail”; Figure 45–2). Therefore, performing a lumbar (subarachnoid) puncture below L1 in an adult (L3 in a child) usually avoids potential needle trauma to the spinal cord; damage to the cauda equina is unlikely, as these nerve roots float in the dural sac below L1 and tend to be pushed away (rather than pierced) by an advancing needle.

A dural sheath invests most nerve roots for a small distance, even after they exit the spinal canal (Figure 45–5). Nerve blocks close to the intervertebral foramen therefore carry a risk of subdural or subarachnoid injection. The dural sac and the subarachnoid and subdural spaces usually extend to S2 in adults and often to S3 in children, important considerations in avoiding accidental dural puncture during caudal anesthesia. An extension of the pia mater, the filum terminale, penetrates the dura and attaches the terminal end of the spinal cord (conus medullaris) to the periosteum of the coccyx (Figure 45–7).

The blood supply to the spinal cord and nerve roots is derived from a single anterior spinal artery and paired posterior spinal arteries (Figure 45–8). The anterior spinal artery is formed from the vertebral artery at the base of the skull and courses down along the anterior surface of the cord. The anterior spinal artery supplies the anterior two-thirds of the cord, whereas the two posterior spinal arteries supply the posterior one-third. The posterior spinal arteries arise from the posterior inferior cerebellar arteries and course down along the dorsal surface of the cord medial to the dorsal nerve roots. The anterior and posterior spinal arteries receive additional blood flow from the intercostal arteries in the thorax and the lumbar arteries in the abdomen. One of these radicular arteries is typically large, the artery of Adamkiewicz, or arteria radicularis magna, arising from the aorta (Figures 45–8A). It is typically unilateral and nearly always arises on the left side, providing the major blood supply to the anterior, lower two-thirds of the spinal cord. Injury to this artery can result in the anterior spinal artery syndrome (see Chapter 22).

The principal site of action for neuraxial blockade is believed to be the nerve root, at least during initial onset of block. Local anesthetics may also have actions on structures within the spinal cord during epidural and spinal anesthesia. Local anesthetic is injected into CSF (spinal anesthesia) or the epidural space (epidural and caudal anesthesia) and bathes the nerve root in the subarachnoid space or epidural space, respectively. Direct injection of local anesthetic into CSF for spinal anesthesia allows a relatively small dose and volume of local anesthetic to achieve dense sensory and motor blockade. In contrast, neuraxial block is achieved only with much larger volumes and quantities of local anesthetic molecules during epidural and caudal anesthesia. The injection site (level) for epidural anesthesia is ideally sited in the middle of the nerve roots that must be anesthetized. Blockade of neural transmission (conduction) in the posterior nerve root fibers interrupts somatic and visceral sensation, whereas blockade of anterior nerve root fibers prevents efferent motor and autonomic outflow.

SOMATIC BLOCKADE

By interrupting the afferent transmission of painful stimuli and abolishing the efferent impulses responsible for skeletal muscle tone, neuraxial blocks provide excellent operating conditions. Sensory blockade interrupts both somatic and visceral painful stimuli. The mechanism of action of local anesthetic agents is discussed in Chapter 16. Smaller and myelinated fibers are generally more easily blocked than larger and unmyelinated ones. The size and character of the fiber types, and the fact that the concentration of local anesthetic decreases with increasing distance from the level of injection, explains the phenomenon of differential blockade during neuraxial anesthesia. Differential blockade typically results in sympathetic blockade (judged by temperature sensitivity) that may be two segments or more cephalad than the sensory block (pain, light touch), which, in turn, is usually several segments more cephalad than the motor blockade.

AUTONOMIC BLOCKADE

Interruption of efferent autonomic transmission at the spinal nerve roots during neuraxial blocks produces sympathetic blockade. Sympathetic outflow from the spinal cord may be described as thoracolumbar, whereas parasympathetic outflow is craniosacral. Sympathetic preganglionic nerve fibers (small, myelinated B fibers) exit the spinal cord with the spinal nerves from T1–L2 and may course many levels up or down the sympathetic chain before synapsing with a postganglionic cell in a sympathetic ganglion. In contrast, parasympathetic preganglionic fibers exit the spinal cord with the cranial and sacral nerves. Neuraxial anesthesia does not block the vagus nerve (tenth cranial nerve). The physiological responses to neuraxial blockade therefore result from decreased sympathetic tone or unopposed parasympathetic tone, or both.

Cardiovascular Manifestations

Neuraxial blocks produce variable decreases in blood pressure that may be accompanied by a decrease in heart rate. These effects generally increase with more cephalad dermatomal levels and more extensive sympathectomy. Vasomotor tone is primarily determined by sympathetic fibers arising from T5 to L1, innervating arterial and venous smooth muscle. Blocking these nerves causes vasodilation of the venous capacitance vessels and pooling of blood in the viscera and lower extremities, thereby decreasing the effective circulating blood volume and often decreasing cardiac output. Arterial vasodilation may also decrease systemic vascular resistance. The effects of arterial vasodilation may be minimized by compensatory vasoconstriction above the level of the block, particularly when the extent of sensory anesthesia is limited to the lower thoracic dermatomes. A high sympathetic block not only prevents compensatory vasoconstriction but may also block the sympathetic cardiac accelerator fibers that arise at T1 to T4. Profound hypotension may result from arterial dilation and venous pooling combined with bradycardia. These effects are exaggerated if venous pooling is further augmented by a head-up position or the weight of a gravid uterus. Unopposed vagal tone may explain the sudden cardiac arrest sometimes seen with spinal anesthesia.

Deleterious cardiovascular effects should be anticipated and steps undertaken to minimize the degree of hypotension. However, volume loading with 10 to 20 mL/kg of intravenous fluid in a healthy patient before initiation of the block has been shown repeatedly to fail to prevent hypotension (in the absence of preexisting hypovolemia). Left uterine displacement in the third trimester of pregnancy helps to minimize physical obstruction to venous return in some patients. Despite these efforts, hypotension may still occur and should be treated promptly. Autotransfusion may be accomplished by placing the patient in a head-down position. A bolus of intravenous fluid (5–10 mL/kg) may be helpful in patients who have adequate cardiac and renal function to be able to “handle” the fluid load after the block wears off. Excessive or symptomatic bradycardia should be treated with atropine, and hypotension should be treated with vasopressors. Direct α-adrenergic agonists (such as phenylephrine) primarily produce vasoconstriction, increasing systemic vascular resistance, and may reflexively increase bradycardia. The “mixed” agent ephedrine has direct and indirect β-adrenergic effects that increase heart rate and contractility, and indirect effects that also produce vasoconstriction. Much like ephedrine, small doses of epinephrine (2–5 mcg boluses) are particularly useful in treating spinal anesthesia–induced hypotension. If profound hypotension or bradycardia persists, vasopressor infusions may be required.

Pulmonary Manifestations

Alterations in pulmonary physiology are usually minimal with neuraxial blocks because the diaphragm is innervated by the phrenic nerve, with fibers originating from C3 to C5. Even with high thoracic levels, tidal volume is unchanged; there is only a small decrease in vital capacity, which results from a loss of the abdominal muscles’ contribution to forced expiration.

Patients with severe chronic lung disease may rely upon accessory muscles of respiration (intercostal and abdominal muscles) to actively inspire or exhale. High levels of neural blockade will impair these muscles. Similarly, effective coughing and clearing of secretions require these muscles for expiration. For these reasons, neuraxial blocks should be used with caution in patients with limited respiratory reserve. These deleterious effects need to be weighed against the advantages of avoiding airway instrumentation and positive-pressure ventilation. For surgical procedures above the umbilicus, a pure regional technique may not be the best choice in patients with severe lung disease. On the other hand, these patients may benefit from the effects of thoracic epidural analgesia (with dilute local anesthetics and opioids) or intrathecal opioids in the postoperative period, particularly following upper abdominal or thoracic surgery. Some evidence suggests that postoperative thoracic epidural analgesia in high-risk patients can improve pulmonary outcome by decreasing the incidence of pneumonia and respiratory failure, improving oxygenation, and decreasing the duration of mechanical ventilatory support.

Gastrointestinal Manifestations

Neuraxial block–induced sympathectomy allows vagal “dominance” with a small, contracted gut and active peristalsis. This can improve operative conditions during intestinal surgery when used as an adjunct to general anesthesia. Postoperative epidural analgesia with local anesthetics and minimal systemic opioids hastens the return of gastrointestinal function after open abdominal procedures.

Hepatic blood flow will decrease with reductions in mean arterial pressure from any anesthetic technique, including neuraxial anesthesia.

Urinary Tract Manifestations

Renal blood flow is maintained through autoregulation, and there is little effect of neuraxial anesthesia on kidney function. Neuraxial anesthesia at the lumbar and sacral levels blocks both sympathetic and parasympathetic control of bladder function. Loss of autonomic bladder control results in urinary retention until the block wears off. If no urinary catheter is placed perioperatively, it is prudent to use the regional anesthetic of shortest duration sufficient for the surgical procedure and to administer the minimal safe volume of intravenous fluid. Patients with urinary retention should be checked for bladder distention after neuraxial anesthesia.

Metabolic & Endocrine Manifestations

Surgical trauma produces a systemic neuroendocrine stress response via activation of somatic and visceral afferent nerve fibers, in addition to a localized inflammatory response. This systemic response includes increased concentrations of adrenocorticotropic hormone, cortisol, epinephrine, norepinephrine, and vasopressin levels, as well as activation of the renin–angiotensin–aldosterone system. Clinical manifestations include intraoperative and postoperative hypertension, tachycardia, hyperglycemia, protein catabolism, suppressed immune responses, and altered renal function. Neuraxial blockade can partially suppress (during major invasive abdominal or thoracic surgery) or totally block (during lower extremity surgery) the neuroendocrine stress response. To maximize this blunting of the neuroendocrine stress response, neuraxial block should precede incision and continue postoperatively.

Indications

Neuraxial blocks may be used alone or in conjunction with general anesthesia for many procedures below the neck. As a primary anesthetic, neuraxial blocks have proved most useful in lower abdominal, inguinal, urogenital, rectal, and lower extremity surgery. Lumbar spinal surgery may also be performed under spinal anesthesia. Upper abdominal procedures (eg, gastrectomy) have been performed with spinal or epidural anesthesia, but because it can be difficult to safely achieve a sensory level adequate for patient comfort, these techniques are less commonly used.

If a neuraxial anesthetic is being considered, the risks and benefits must be discussed with the patient, and informed consent should be obtained. The patient must be mentally prepared for neuraxial anesthesia, and neuraxial anesthesia must be appropriate for the type of surgery. Patients should understand that they will have little or no lower extremity motor function until the block resolves. Procedures that require maneuvers that might compromise respiratory function (eg, pneumoperitoneum or pneumothorax) or those operations that are unusually of long duration are typically performed with general anesthesia, with or without neuraxial blockade.

Contraindications

Major contraindications to neuraxial anesthesia include lack of consent, coagulation abnormalities, severe hypovolemia, elevated intracranial pressure (particularly with an intracranial mass), and infection at the site of injection. Other relative contraindications include severe aortic or mitral stenosis and severe left ventricular outflow obstruction (hypertrophic obstructive cardiomyopathy); however, with close monitoring and control of the anesthetic level, neuraxial anesthesia can be performed safely in patients with stenotic valvular heart disease, particularly if extensive dermatomal spread of anesthesia is not required (eg, “saddle” block spinal anesthetics).

Relative and controversial contraindications are also shown in Table 45–1. Inspection and palpation of the back can reveal surgical scars, scoliosis, skin lesions, and whether or not the spinous processes can be identified. Although preoperative screening tests are not required in healthy patients undergoing neuraxial blockade, appropriate testing should be performed if the clinical history suggests a coagulation abnormality. Neuraxial anesthesia in the presence of sepsis or bacteremia could theoretically predispose patients to hematogenous spread of the infectious agents into the epidural or subarachnoid space.

Patients with preexisting neurological deficits or demyelinating diseases may report worsening symptoms following a neuraxial block. It may be impossible to discern effects or complications of the block from preexisting deficits or unrelated exacerbation of preexisting disease. For these reasons, some risk-averse practitioners argue against neuraxial anesthesia in such patients. A preoperative neurological examination should thoroughly document any deficits. In a retrospective study examining the records of 567 patients with preexisting neuropathies, 2 of the patients developed new or worsening neuropathy following neuraxial anesthesia. Although this finding indicates a relatively low risk of further injury, study investigators suggest that an injured nerve is vulnerable to additional injury, increasing the likelihood of poor neurological outcomes. However, a history of preexisting neurological deficits or demyelinating disease is at best a relative contraindication, and the balance of perioperative risks in this patient population may favor neuraxial anesthesia in certain select patients.

Regional anesthesia requires at least some degree of patient cooperation. This may be difficult or impossible for patients with dementia, psychosis, or emotional instability. The decision must be individualized. Unsedated young children may not be suitable for pure regional techniques; however, regional anesthesia is frequently used with general anesthesia in children.

Neuraxial Blockade in the Setting of Anticoagulants & Antiplatelet Agents

Whether a block should be performed in the setting of anticoagulants and antiplatelet agents can be problematic. The American Society of Regional Anesthesia and Pain Medicine (ASRA) has issued guidelines on this subject. Because guidelines are frequently revised and updated, practitioners are advised to seek the most recent edition. Fortunately, the incidence of epidural hematoma is reported to be infrequent (1 in 150,000 epidurals). The use of anticoagulant and antiplatelet medications continues to increase, placing an ever-larger number of patients at potential risk of epidural hematomas. However, because of the rarity of epidural hematomas, most guidelines are based on expert opinion and case series reviews, as clinical trials are not feasible.

A. Oral Anticoagulants

If neuraxial anesthesia is to be used in patients receiving warfarin therapy, a normal prothrombin time and international normalized ratio usually will be documented prior to the block, unless the drug has been discontinued for weeks. Anesthesia staff should always consult with the patient’s primary physicians whenever considering the discontinuation of antiplatelet or antithrombotic therapy. New agents such as the direct thrombin inhibitor dabigatran and the factor X_a inhibitors rivaroxaban and apixaban are increasingly encountered by anesthesia staff (Figure 45–9).

Two drug half-life intervals have been suggested as the time from drug discontinuation until neuraxial procedures are performed. However, depending upon patient factors two half-life intervals may not be sufficient to mitigate the increased risk of bleeding following neuraxial procedures. A period from drug discontinuation of up to five to six drug half-life intervals may be necessary to avoid increased bleeding risk. Thrombin clotting time assays can be used to detect the effects of dabigatran. Likewise, factor X_a inhibitors can be assessed through assays of factor X_a inhibition. Anesthesia providers planning neuraxial procedures should consult closely with the patient’s primary providers to discern if suspension of anticoagulation is advised when considering a neuraxial technique. Bridging therapy with low-molecular-weight heparin can be considered during the time that oral anticoagulation is suspended if there is increased thrombotic risk. It has been suggested that new oral anticoagulants can be resumed 24 to 48 h following a neuraxial procedure or removal of an epidural catheter.

B. Antiplatelet Drugs

By themselves, aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) drugs do not increase the risk of spinal hematoma from neuraxial anesthesia procedures or epidural catheter removal. This assumes a normal patient with a normal coagulation profile who is not receiving other medications that might affect clotting mechanisms. In contrast, more potent agents should be stopped, and neuraxial blockade should generally be administered only after their effects have worn off. The waiting period depends on the specific agent: for ticlopidine it is 14 days; clopidogrel, 7 days; prasugrel, 7 to 10 days; ticagrelor, 5 days; abciximab, 48 h; and eptifibatide, 8 h. Neuraxial techniques should be avoided in patients receiving antiplatelet medications until platelet function has been recovered. Metabolites of clopidogrel and prasugrel block the P2Y12 receptor, impeding platelet aggregation. Ticagrelor directly inhibits the P2Y12 receptor. Both prasugrel and ticagrelor have greater platelet inhibition compared with clopidogrel. In patients with a recently placed cardiac stent, discontinuation of antiplatelet therapy can result in stent thrombosis and acute ST-segment elevation myocardial infarction. Risks versus benefits of a neuraxial technique should be discussed with the patient and the patient’s primary physicians.

C. Standard (Unfractionated) Heparin

“Minidose” subcutaneous heparin prophylaxis is not a contraindication to neuraxial anesthesia or epidural catheter removal. In patients who are to receive systemic heparin intraoperatively, blocks may be performed 1 h or more before heparin administration. A bloody epidural or spinal catheter placement does not necessarily require cancellation of surgery, but discussion of the risks with the surgeon and careful postoperative monitoring is needed. Removal of an epidural catheter should occur 1 h prior to, or 4 h following, subsequent heparin dosing.

Neuraxial anesthesia should be avoided in patients on therapeutic doses of intravenous heparin and with increased partial thromboplastin time. If the patient is started on heparin after the placement of an epidural catheter, the catheter should be removed only after discontinuation or interruption of heparin infusion and evaluation of the coagulation status. The risk of spinal hematoma is unclear in the setting of the anticoagulation required for cardiac surgery, and we have used epidural techniques for pain control after such procedures as off-pump coronary bypass only in the rare instances where the anticipated benefits exceed the perceived risks. Prompt diagnosis and evacuation of symptomatic epidural hematomas increase the likelihood that neuronal function will be preserved.

D. Low-Molecular-Weight Heparin (LMWH)

A number of cases of spinal hematoma were reported associated with neuraxial anesthesia after patients began receiving low-molecular-weight heparin (LMWH) enoxaparin (Lovenox) in the United States in 1993. Many of these cases involved intraoperative or early postoperative LMWH use, and several patients were receiving concomitant antiplatelet medication. The ASRA guideline provides useful recommendations for risk reduction in such circumstances. Some of the draft recommendations include that if an unusually bloody needle or catheter placement occurs, LMWH should be delayed until 24 h postoperatively, because this trauma may increase the risk of spinal hematoma. If postoperative LMWH thromboprophylaxis will be utilized, epidural catheters should be removed 4 h prior to the first LMWH dose.

E. Fibrinolytic or Thrombolytic Therapy

Neuraxial anesthesia should not be performed if a patient has received fibrinolytic or thrombolytic therapy.

ASRA has developed a smartphone application (ASRA Coags Regional app) to assist in the perioperative management of patients taking drugs that affect coagulation.

Awake or Asleep?

Should lumbar neuraxial anesthesia, when used in conjunction with general anesthesia, be performed after induction of general anesthesia? This is controversial. The major arguments for having the patient asleep are that (1) most patients, if given a choice, would prefer to be asleep, and (2) the possibility of sudden patient movement causing injury is markedly diminished. The major argument in favor of neuraxial blockade administration only while the patient is still awake is that the patient can alert the clinician to paresthesias and pain on injection in this circumstance, both of which have been associated with postoperative neurological deficits. Although many clinicians are comfortable performing lumbar epidural or spinal puncture in anesthetized or deeply sedated adults, there is greater consensus, although not unanimous opinion, that thoracic and cervical punctures should, except under unusual circumstances, only be performed in awake, responsive patients. Pediatric neuraxial blocks, particularly caudal and epidural blocks, are usually performed under general anesthesia.

Technical Considerations

Neuraxial blocks must be performed only in a facility in which all the equipment and drugs needed for intubation, resuscitation, and general anesthesia are immediately available. Regional anesthesia is greatly facilitated by adequate patient premedication. Nonpharmacological patient preparation is also very helpful. The patient should be told what to expect so as to minimize anxiety. This is particularly important in situations in which premedication is not used, as is typically the case in obstetric anesthesia. Supplemental oxygen via a face mask or nasal cannula may be required to avoid hypoxemia when sedation is used. Minimum monitoring requirements include blood pressure and pulse oximetry for labor analgesia. Monitoring for blocks rendered in surgical anesthesia is the same as that in general anesthesia.

Surface Anatomy

Spinous processes are usually palpable and help to define the midline. Ultrasound can be used when landmarks are not palpable (Figure 45–10). The spinous processes of the cervical and lumbar spine are nearly horizontal, whereas those in the thoracic spine slant in a caudal direction and can overlap significantly (Figure 45–2). Therefore, when performing a lumbar or cervical epidural block (with maximum spinal flexion), the needle is directed with only a slight cephalad angle, if at all, whereas for a thoracic block, the needle must be angled significantly more cephalad to enter the thoracic epidural space. In the cervical area, the first palpable spinous process is that of C2, but the most prominent one is that of C7 (vertebra prominens). With the arms at the side, the spinous process of T7 is usually at the same level as the inferior angle of the scapulae (Figure 45–11). A line drawn between the highest points of both iliac crests (Tuffier’s line) usually crosses either the body of L4 or the L4–L5 interspace. Counting spinous processes up or down from these reference points identifies other spinal levels. A line connecting the posterior superior iliac spine crosses the S2 posterior foramina. In slender persons, the sacrum is easily palpable, and the sacral hiatus is felt as a depression just above or between the gluteal clefts and above the coccyx, defining the point of entry for caudal blocks.

Patient Positioning

A. Sitting Position

The anatomic midline is often easier to identify when the patient is sitting than when the patient is in the lateral decubitus position (Figure 45–12). This is particularly true with obese patients. Patients sit with their elbows resting on their thighs or a bedside table, or they can hug a pillow. Flexion of the spine (arching the back “like an angry cat”) maximizes the “target” area between adjacent spinous processes and brings the spine closer to the skin surface (Figure 45–13).

B. Lateral Decubitus

Many clinicians prefer the lateral position for neuraxial blocks (Figure 45–14). Patients lie on their side with their knees flexed and pulled high against the abdomen or chest, assuming a “fetal position.” An assistant can help the patient assume and hold this position.

C. Buie’s (Jackknife) Position

This position may be used for anorectal procedures utilizing an isobaric or hypobaric anesthetic solution (see later discussion). The advantage is that the block is done in the same position as the operative procedure, so that the patient does not have to be moved following the block. The disadvantage is that CSF will not freely flow through the needle, so that correct subarachnoid needle tip placement will need to be confirmed by CSF aspiration. A prone position is typically used when fluoroscopic guidance is required.

Anatomic Approach

A. Midline Approach

The spine is palpated, and the patient’s body positioned so that a needle passed parallel to the floor will stay midline as it courses deeper (Figure 45–4). The depression between the spinous processes of the vertebrae above and below the level to be used is palpated; this will be the needle entry site. A sterile field is established with an appropriate antibacterial solution. A fenestrated sterile drape is applied. After the preparation solution has dried, a skin wheal is raised at the level of the chosen interspace with local anesthetic using a small (25-gauge) needle. A longer needle can be used for deeper local anesthetic infiltration.

Next, the procedure needle is introduced in the midline. Remembering that the spinous processes course caudad from their origin at the spine, the needle will be directed slightly cephalad. The subcutaneous tissues offer little resistance to the needle. As the needle courses deeper, it will enter the supraspinous and interspinous ligaments, felt as an increase in tissue resistance. The needle also feels more firmly implanted in the back (like “an arrow in a target”). If bone is contacted superficially, a midline needle is likely hitting the lower spinous process. Contact with bone at a deeper level usually indicates that the needle is in the midline and hitting the upper spinous process, or that it is lateral to the midline and hitting a lamina. In either case, the needle must be redirected. As the needle penetrates the ligamentum flavum, an obvious increase in resistance is encountered. At this point, the procedures for spinal and epidural anesthesia differ.

For epidural anesthesia, a sudden loss of resistance (to injection of air or saline) is encountered as the needle passes through the ligamentum flavum and enters the epidural space. For spinal anesthesia, the needle is advanced through the epidural space and penetrates the dura–subarachnoid membranes, as signaled by freely flowing CSF.

B. Paramedian Approach

The paramedian technique may be selected, particularly if epidural or subarachnoid block is difficult, particularly in patients who cannot be positioned easily (eg, severe arthritis, kyphoscoliosis, or prior spine surgery) (Figure 45–15). Many clinicians routinely use the paramedian approach for thoracic epidural puncture. After skin preparation and sterile draping (as previously described), the skin wheal for a paramedian approach is raised 2 cm lateral to the inferior aspect of the superior spinous process of the desired level. Because this approach is lateral to most of the interspinous ligaments and penetrates the paraspinous muscles, the needle may encounter little resistance initially and may not seem to be in firm tissue. The needle is directed and advanced at a 10° to 25° angle toward the midline. If bone is encountered at a shallow depth with the paramedian approach, the needle is likely in contact with the medial part of the lower lamina and should be redirected mostly upward and perhaps slightly more laterally. On the other hand, if bone is encountered deeply, the needle is usually in contact with the lateral part of the lower lamina and should be redirected only slightly craniad, more toward the midline (Figure 45–16).

C. Assessing Level of Blockade

With knowledge of the sensory dermatomes (see appendix), the extent of sensory block can be assessed by a blunted needle or a piece of ice.

D. Ultrasound-Guided Neuraxial Blockade

Although it has not, as of yet, transformed the practice of neuraxial blockade in the same manner as it has for other procedures, ultrasound guidance can facilitate neuraxial blockade in patients with poorly palpable landmarks. As with other uses of ultrasound, specific training is required for practitioners to identify correctly the landmarks and interspaces necessary for neuraxial blockade.

Spinal Needles

Spinal needles are commercially available in an array of sizes lengths, and bevel and tip designs (Figure 45–17). All should have a tightly fitting, removable stylet that completely occludes the lumen to avoid tracking epithelial cells into the subarachnoid space. Broadly, they can be divided into either sharp (cutting)-tipped or blunt-tipped needles. The Quincke needle is a cutting needle with end injection. The introduction of blunt tip (pencil-point) needles has markedly decreased the incidence of postdural puncture headache. The Whitacre and other pencil-point needles have rounded points and side injection. The Sprotte is a side-injection needle with a long opening. It has the advantage of more vigorous CSF flow compared with similar gauge needles. However, this can lead to a failed block if the distal part of the opening is subarachnoid (with free flow CSF), the proximal part is not past the dura, and the full dose of medication is not delivered. In general, the smaller the gauge needle (along with use of a blunt-tipped needle), the lower will be the incidence of headache.

Spinal Catheters

Larger catheters designed for epidural use are frequently employed for continuous spinal anesthesia following accidental dural puncture during performance of epidural anesthesia. Catheters must be carefully labeled as being subarachnoid, as opposed to epidural, to avoid the potential for inadvertent dosage.

Specific Technique for Spinal Anesthesia

The midline or paramedian approaches, with the patient positioned in the lateral decubitus, sitting, or prone positions, can be used for spinal anesthesia. As previously discussed, the needle is advanced from skin through the deeper structures until two “pops” are felt. The first is penetration of the ligamentum flavum, and the second is penetration of the dura–arachnoid membrane. Successful dural puncture is confirmed by withdrawing the stylet to verify free flow of CSF. With small-gauge needles (<25 gauge), aspiration may be necessary to detect CSF. If free flow occurs initially, but CSF cannot be aspirated after attaching the syringe, the needle likely will have moved. Persistent paresthesias or pain with injection of drugs should prompt the clinician to withdraw and redirect the needle.

Factors Influencing Level of Spinal Block

Table 45–2 lists factors that have been shown to affect the level of neural blockade following spinal anesthesia. The most important determinants are baricity of the local anesthetic solution, position of the patient during and immediately after injection, and drug dosage. In general, the larger the dosage or more cephalad the site of injection, the more cephalad the level of anesthesia that will be obtained. Moreover, migration of the local anesthetic cephalad in CSF depends on its density relative to CSF (baricity). CSF has a specific gravity of 1.003 to 1.008 at 37°C. Table 45–3 lists the specific gravity of anesthetic solutions. A hyperbaric solution of local anesthetic is denser (heavier) than CSF, whereas a hypobaric solution is less dense (lighter) than CSF. The local anesthetic solutions can be made hyperbaric by the addition of glucose or hypobaric by the addition of sterile water or fentanyl. Thus, with the patient in a head-down position, a hyperbaric solution spreads cephalad, and a hypobaric anesthetic solution moves caudad. A head-up position causes a hyperbaric solution to settle caudad and a hypobaric solution to ascend cephalad. Similarly, when a patient remains in a lateral position, a hyperbaric spinal solution will have a greater effect on the dependent (down) side, whereas a hypobaric solution will achieve a higher level on the nondependent (up) side. An isobaric solution tends to remain at the level of injection. Anesthetic agents lacking glucose may be mixed with CSF (at least 1:1) to make their solutions isobaric. Other factors affecting the level of neural blockade include the level of injection and the patient’s height and vertebral column anatomy. The direction of the needle bevel or injection port may also play a role; higher levels of anesthesia are achieved if the injection is directed cephalad than if the point of injection is oriented laterally or caudad.

Hyperbaric solutions tend to move to the most dependent area of the spine (normally T4–T8 in the supine position).

With normal spinal anatomy, the apex of the thoracolumbar curvature is T4 (Figure 45–18). In the supine position, this should limit a hyperbaric solution to produce a level of anesthesia at or below T4. Abnormal curvatures of the spine, such as scoliosis and kyphoscoliosis, have multiple effects on spinal anesthesia. Placing the block becomes more difficult because of the rotation and angulation of the vertebral bodies and spinous processes. Finding the midline and the interlaminar space may be difficult. The paramedian approach to lumbar puncture may be preferable in patients with severe scoliosis and kyphoscoliosis. Reviewing radiographs of the spine before attempting the block may be useful. Spinal curvature affects the ultimate level by changing the contour of the subarachnoid space. Previous spinal surgery can similarly result in technical difficulties in placing a block. Correctly identifying the interspinous and interlaminar spaces may be difficult at the levels of previous laminectomy or spinal fusion. The paramedian approach may be easier, or a level above the surgical site can be chosen. The block may be incomplete, or the level may be different than anticipated, due to postsurgical anatomic changes.

Lumbar CSF volume inversely correlates with the dermatomal spread of spinal anesthesia. Increased intraabdominal pressure or other conditions that cause engorgement of the epidural veins, thus decreasing CSF volume, are associated with greater dermatomal spread for a given volume of injectate. This would include conditions such as pregnancy, ascites, and large abdominal tumors. In these clinical situations, higher levels of anesthesia are achieved with a given dose of local anesthetic than would otherwise be expected. For spinal anesthesia on a term parturient, some clinicians reduce the dosage of anesthetic by one-third compared with a nonpregnant patient, particularly when the block will be initiated with the patient in the lateral position. Age-related decreases in CSF volume are likely responsible for the higher anesthetic levels achieved in the elderly for a given dosage of spinal anesthetic. Severe kyphosis or kyphoscoliosis can also be associated with a decreased volume of CSF and often results in a higher than expected level, particularly with a hypobaric technique or rapid injection.

Spinal Anesthetic Agents

Many local anesthetics have been used for spinal anesthesia in the past, but only a few are currently in use (Table 45–4). Only preservative-free local anesthetic solutions are used. Addition of vasoconstrictors (α-adrenergic agonists, epinephrine [0.1–0.2 mg]) and opioids enhance the quality or prolong the duration of spinal anesthesia, or both. Vasoconstrictors seem to delay the uptake of local anesthetics from CSF and may have weak spinal analgesic properties. Opioids and clonidine can likewise be added to spinal anesthetics to improve both the quality and duration of the subarachnoid block.

Until very recently in North America, hyperbaric spinal anesthesia was more commonly used than hypobaric or isobaric techniques. The level of anesthesia is then dependent on the patient’s position during and immediately following the injection. In the sitting position, “saddle block” can be achieved by keeping the patient sitting for 3 to 5 min following injection, so that only the lower lumbar nerves and sacral nerves are blocked. If the patient is moved from a sitting position to a supine position immediately after injection, the agent will move more cephalad to the dependent region defined by the thoracolumbar curve. Hyperbaric anesthetics injected intrathecally with the patient in a lateral decubitus position are useful for unilateral lower extremity procedures. The patient is placed laterally, with the extremity to be operated on in a dependent position. If the patient is kept in this position for about 5 min following injection, the block will tend to be denser and achieve a higher level on the operative dependent side.

If regional anesthesia is chosen for surgical procedures involving hip or lower extremity fracture, hypobaric or isobaric spinal anesthesia can be useful because the patient need not lie on the fractured extremity.

Continuous epidural anesthesia is a neuraxial technique offering a range of applications wider than single-dose spinal anesthesia. An epidural block can be performed at the lumbar, thoracic, or cervical level. Sacral epidural anesthesia is referred to as a caudal block and is described at the end of this chapter. Epidural techniques are widely used for surgical anesthesia, obstetric analgesia, postoperative pain control, and chronic pain management. Epidurals can be used as a single shot technique or with a catheter that allows intermittent boluses or continuous infusion, or both. The motor block can range from none to complete. All of these variables are controlled by the choice of drug, concentration, dosage, and level of injection.

The epidural space surrounds the dura mater posteriorly, laterally, and anteriorly. Nerve roots travel in this space as they exit laterally through the foramen and course outward to become peripheral nerves. Other contents of the lumbar epidural space include fatty connective tissue, lymphatics, and a rich venous (Batson’s) plexus. Fluoroscopic studies have demonstrated the presence of septa or connective tissue bands within the epidural space, possibly explaining the occasional one-sided epidural block. Epidural anesthesia is slower in onset (10–20 min) and may not be as dense as spinal anesthesia, a feature that can be useful clinically. For example, by using relatively dilute concentrations of a local anesthetic combined with an opioid, an epidural provides analgesia without motor block. This is commonly employed for labor and postoperative analgesia. Moreover, a segmental block is possible because the anesthetic can be confined close to the level at which it was injected. A segmental block is characterized by a well-defined band of anesthesia at certain nerve roots; leaving nerve roots above and below unblocked. This can be seen with a thoracic epidural that provides upper abdominal anesthesia while sparing cervical and lumbar nerve roots.

Epidural anesthesia is most often performed in the lumbar region. The midline (Figure 45–4) or paramedian approaches (Figure 45–15) can be used. Lumbar epidural anesthesia can be used for any procedure below the diaphragm. Because the spinal cord typically terminates at the L1 level, there is an extra measure of safety in performing the block in the lower lumbar interspaces, particularly if an accidental dural puncture occurs (see “Complications,” later).

Thoracic epidural blocks are technically more difficult to accomplish than are lumbar blocks because of greater angulation and the overlapping of the spinous processes at the vertebral level (Figure 45–19). Moreover, the potential risk of spinal cord injury with accidental dural puncture, although exceedingly small with good technique, may be greater than that at the lumbar level. Thoracic epidural blocks can be accomplished with either a midline or paramedian approach. The thoracic epidural technique is more commonly used for postoperative analgesia than as a primary anesthetic. Single shot or catheter techniques are used for the management of chronic pain. Infusions via an epidural catheter are useful for providing prolonged durations of analgesia and may obviate or shorten postoperative ventilation in patients with underlying lung disease and following chest surgery.

Cervical blocks are usually performed with the patient sitting, with the neck flexed, using the midline approach. They are used most often for the management of acute and chronic pain.

Epidural Needles

The standard epidural needle is typically 17 to 18 gauge, 3 or 3.5 inches long, and has a blunt bevel with a gentle curve of 15° to 30° at the tip. The Tuohy needle is most commonly used (Figure 45–20). The blunt, curved tip theoretically helps to push away the dura after passing through the ligamentum flavum instead of penetrating it. Straight needles without a curved tip (Crawford needles) may have a greater incidence of dural puncture. Needle modifications include winged tips and introducer devices set into the hub designed for guiding catheter placement.

Epidural Catheters

Placing a catheter into the epidural space allows for continuous infusion or intermittent bolus techniques. In addition to extending the duration of the block, catheter techniques may allow a lower total dose of anesthetic to be used.

Epidural catheters are useful for intraoperative epidural anesthesia and postoperative analgesia. Typically, a 19- or 20-gauge catheter is introduced through a 17- or 18-gauge epidural needle. When using a curved tipped needle, the bevel opening is directed either cephalad or caudad, and the catheter is advanced 2 to 6 cm into the epidural space. The shorter the distance the catheter is advanced, the more likely it is to become dislodged. Conversely, the further the catheter is advanced, the greater the chance of a unilateral block (due to the catheter tip either exiting the epidural space via an intervertebral foramen or coursing into the anterolateral recesses of the epidural space) and the greater chance of penetration of an epidural vein occurring. After advancing the catheter the desired depth, the needle is removed, leaving the catheter in place. The catheter can be taped or otherwise secured along the back. Catheters that will remain in place for prolonged times (eg, >1 week) may be tunneled under the skin. Catheters have either a single port at the distal end or multiple side ports close to a closed tip. Some have a stylet for easier insertion or to help steer the catheter passage in the epidural space with fluoroscopic guidance. Spiral wire-reinforced catheters are very resistant to kinking. The spiral or spring tip is associated with fewer, less intense paresthesias and may be associated with a lower incidence of inadvertent intravascular insertion.

Specific Techniques for Epidural Anesthesia

Using the midline or paramedian approaches detailed previously, the epidural needle is passed through the skin and the ligamentum flavum. The needle must stop short of piercing the dura. Two techniques make it possible to determine when the tip of the needle has entered the potential (epidural) space: the “loss of resistance” and “hanging drop” techniques.

The loss of resistance technique is preferred by most clinicians. The needle is advanced through the subcutaneous tissues with the stylet in place until the interspinous ligament is entered, as noted by an increase in tissue resistance. The stylet or introducer is removed, and a glass syringe filled with saline or air is attached to the hub of the needle. If the tip of the needle is within the ligament, gentle attempts at injection are met with resistance, and injection is not possible. The needle is then slowly advanced, millimeter by millimeter, with either continuous or rapidly repeating attempts at injection. As the tip of the needle just enters the epidural space, there is a sudden loss of resistance, and injection is easy.

Once the interspinous ligament has been entered and the stylet has been removed, the hanging drop technique requires that the hub of the needle be filled with solution so that a drop hangs from its outside opening. The needle is then slowly advanced deeper. As long as the tip of the needle remains within the ligamentous structures, the drop remains “hanging.” However, as the tip of the needle enters the epidural space, it creates negative pressure, and the drop of fluid is sucked into the needle. If the needle becomes plugged, the drop will not be drawn into the hub of the needle, and accidental dural puncture may occur. Some clinicians prefer to use this technique for the paramedian approach and for cervical epidurals. Skilled “epiduralists” rely on either the loss of resistance or hanging drop as confirmation (rather than as the primary test) that the needle has entered the epidural space. Successful “epiduralists” will generally have sensed the “give” in their hands as the epidural needle tip passes through the ligamentum flavum.

Activating an Epidural

The quantity (volume and concentration) of local anesthetic needed for epidural anesthesia is larger than that needed for spinal anesthesia. Toxic side effects are almost guaranteed if a “full epidural dose” is injected intrathecally or intravascularly. Safeguards against toxic epidural side effects include test dosing and incremental dosing. These safeguards apply whether the injection is through the needle or through an epidural catheter.

A test dose is designed to detect both subarachnoid and intravascular injection. The classic test dose combines local anesthetic and epinephrine, typically 3 mL of 1.5% lidocaine with 1:200,000 epinephrine (0.005 mg/mL). The 45 mg of lidocaine, if injected intrathecally, will produce spinal anesthesia that should be rapidly apparent. Some clinicians have suggested the use of lower doses of local anesthetic, as an unintended injection of 45 mg of intrathecal lidocaine, can be difficult to manage in areas such as labor rooms. The 15 mcg dose of epinephrine, if injected intravascularly, should produce a noticeable increase in heart rate (20% or more), with or without hypertension. Unfortunately, epinephrine as a marker of intravenous injection is not ideal. False positives (a uterine contraction causing pain or an increase in heart rate coincident to test dosing) and false negatives (bradycardia and exaggerated hypertension in response to epinephrine in patients taking β-blockers) can occur. Simply aspirating prior to injection is insufficient to avoid inadvertent intravenous injection; most experienced practitioners have encountered false-negative aspirations through both a needle and a catheter.

Incremental dosing is a very effective method of avoiding serious complications (“each dose is a test dose”). If aspiration is negative, a fraction of the total intended local anesthetic dose is injected, typically 5 mL. This dose should be large enough for mild symptoms (tinnitus or metallic taste) or signs (slurred speech, altered mentation) of intravascular injection to occur, but small enough to avoid seizure or cardiovascular compromise. This is particularly important for labor epidurals that are to be used for cesarean delivery. If the initial labor epidural bolus was delivered through the needle, and the catheter was then inserted, it may be erroneously assumed that the catheter is well positioned because the patient is still comfortable from the initial bolus. If the catheter was inserted into a blood vessel, or after initial successful placement, has since migrated intravascularly, systemic toxicity will likely result if the full anesthetic dose is injected. Catheters can migrate intrathecally or intravascularly from an initially correct epidural position at any time after placement, but migration is most likely to occur with movement of the patient.

If a clinician uses an initial test dose, is diligent about aspirating prior to each injection, and always uses incremental dosing, major systemic toxic side effects and total spinal anesthesia from accidental intrathecal injections will be rare. Rescue lipid emulsion (20% Intralipid 1.5 mL/kg) must be available whenever epidural blocks are performed, in the event of local anesthetic systemic toxicity.

Factors Affecting Level of Block

Factors affecting the level of epidural anesthesia may not be as predictable as with spinal anesthesia. In adults, 1 to 2 mL of local anesthetic per segment to be blocked is a generally accepted guideline. For example, to achieve a T4 sensory level from an L4–L5 injection would require about 12 to 24 mL. For segmental or analgesic blocks, less volume is needed.

The dose required to achieve the same level of anesthesia decreases with age. This is probably a result of age-related decreases in the size or compliance of the epidural space. Although there is little correlation between body weight and epidural dosage requirements, patient height affects the extent of cephalad spread. Thus, shorter patients may require only 1 mL of local anesthetic per segment to be blocked, whereas taller patients generally require 2 mL per segment. Although less dramatic than with hyperbaric or hypobaric spinal anesthesia, spread of epidural local anesthetics tends to be partially affected by gravity. The lateral decubitus, Trendelenburg, and reverse Trendelenburg positions can be used to help achieve blockade in the desired dermatomes.

Additives to the local anesthetic, particularly opioids, tend to have a greater effect on the quality of epidural anesthesia than on the duration of the block. Epinephrine in concentrations of 5 mcg/mL prolongs the effect of epidural lidocaine, mepivacaine, and chloroprocaine more than that of bupivacaine, levobupivacaine, or ropivacaine. In addition to prolonging the duration and improving the quality of block, epinephrine delays vascular absorption and reduces peak systemic blood levels of all epidurally administered local anesthetics.

Epidural Anesthetic Agents

The epidural agent is chosen based on the desired clinical effect, whether it is to be used as a primary anesthetic, supplementation of general anesthesia, or analgesia. The anticipated duration of the procedure may call for a short- or long-acting single shot anesthetic or the insertion of a catheter (Table 45–5). Commonly used short- to intermediate-acting agents for surgical anesthesia include chloroprocaine, lidocaine, and mepivacaine. Longer acting agents include bupivacaine, levobupivacaine, and ropivacaine.

Following the initial 1 to 2 mL per segment bolus (in fractionated doses), repeat doses delivered through an epidural catheter are done on a fixed time interval (either as a bolus or continuous infusion), based on the practitioner’s experience with the agent, or only redosed when the block demonstrates some degree of regression. Once some regression in sensory level has occurred, one-third to one-half of the initial activation dose can generally safely be reinjected in incremental doses.

It should be noted that chloroprocaine, an ester with rapid onset, short duration, and extremely low toxicity, may interfere with the analgesic effects of epidural opioids. Previous chloroprocaine formulations with preservatives, specifically bisulfite and ethylenediaminetetraacetic acid (EDTA), produced cauda equine syndrome when accidentally injected in a large volume intrathecally. Bisulfite preparations of chloroprocaine were believed to be associated with neurotoxicity, whereas EDTA formulations were associated with severe back pain (presumably due to localized hypocalcemia). Most current preparations of chloroprocaine are preservative-free and without these complications.

Surgical anesthesia is obtained with a 0.5% bupivacaine formulation. The 0.75% formulation of bupivacaine is no longer used in obstetrics, as its use in cesarean delivery has been associated with reports of cardiac arrest after accidental intravenous injection Very dilute concentrations of bupivacaine (eg, 0.0625%) are commonly combined with fentanyl and used for analgesia for labor and postoperative pain. Compared with bupivacaine, ropivacaine may produce less motor block at similar concentrations while maintaining a satisfactory sensory block.

Local Anesthetic pH Adjustment

Local anesthetic solutions have an acidic pH for chemical stability and bacteriostasis. Local anesthetic solutions that are formulated with epinephrine by the manufacturer are more acidic than the “plain” solutions that do not contain epinephrine. Because they are weak bases, they exist primarily in the ionic form in commercial preparations. The onset of neural block requires permeation of lipid barriers by the uncharged form of the local anesthetic. Increasing the pH of the solutions increases the fraction of the uncharged form of the local anesthetic. Addition of sodium bicarbonate (1 mEq/10 mL of local anesthetic) immediately before injection may therefore accelerate the onset of the neural blockade. Sodium bicarbonate is typically not added to bupivacaine, which precipitates above a pH of 6.8.

Failed Epidural Blocks

Unlike spinal anesthesia, in which the procedural endpoint is usually very clear (free-flowing CSF), the onset is very fast, and the technique has a very high success rate, epidural anesthesia is dependent on detection of a more subjective loss of resistance (or hanging drop). Also, the onset of epidural anesthesia is slower, and the more variable anatomy of the epidural space and less predictable spread of local anesthetic make epidural anesthesia inherently less predictable than spinal anesthesia.

Misplaced injections of local anesthetic can occur in a number of situations. In some patients, the spinal ligaments are soft, and either good resistance is never appreciated or a false loss of resistance is encountered. Similarly, entry into the paraspinous muscles during an off-center midline approach may cause a false loss of resistance. Other causes of failed epidural anesthesia (such as intrathecal, subdural, and intravenous injection) are discussed in the late section of this chapter on complications.

A unilateral block can occur if the medication is delivered through a catheter that has either exited the epidural space or coursed laterally. The chance of this occurring increases as longer lengths of catheter are threaded into the epidural space. When unilateral block occurs, the problem may be overcome by withdrawing the catheter 1 to 2 cm and reinjecting it with the patient turned with the unblocked side down. Segmental sparing, which may be due to septations within the epidural space, may also be corrected by injecting additional local anesthetic with the unblocked segment positioned down. The large size of the L5, S1, and S2 nerve roots may delay adequate penetration of local anesthetic and is thought to be responsible for sacral sparing. The latter is particularly a problem for surgery on the knee, ankle, or foot; in such cases, elevating the head of the bed and reinjecting the catheter with additional anesthetic solution can sometimes achieve a more intense block of these large nerve roots. Patients may complain of visceral pain despite a seemingly good epidural block. In some cases (eg, traction on the inguinal ligament and spermatic cord), a high thoracic sensory level may alleviate the pain; in other cases (traction on the peritoneum), intravenous supplementation with opioids or other agents may be necessary. Visceral afferent fibers that travel with the vagus nerve may be responsible.

Caudal epidural anesthesia is a common regional technique in pediatric patients. It may also be used for anorectal surgery in adults. The caudal space is the sacral portion of the epidural space. Caudal anesthesia involves needle or catheter penetration of the sacrococcygeal ligament covering the sacral hiatus that is created by the unfused S4 and S5 laminae. The hiatus may be felt as a groove or notch above the coccyx and between two bony prominences, the sacral cornua (Figure 45–3). Its anatomy is very easily appreciated in infants and children (Figure 45–21). The posterior superior iliac spines and the sacral hiatus define an equilateral triangle (Figure 45–14). Calcification of the sacrococcygeal ligament may make caudal anesthesia difficult or impossible in older adults. As previously noted, the dural sac extends to the first sacral vertebra in adults and to about the third sacral vertebra in infants, making accidental intrathecal injection a common concern in infants.

In children, caudal anesthesia is typically combined with general anesthesia for intraoperative supplementation and postoperative analgesia. It is commonly used for procedures below the diaphragm, including urogenital, rectal, inguinal, and lower extremity surgery. Pediatric caudal blocks are most commonly performed after the induction of general anesthesia. However, regional techniques are increasingly used for surgical anesthesia in infants and young children because of concerns about the possible neurotoxic effects of general anesthesia in that population. The patient is placed in the lateral or prone position with one or both hips flexed, and the sacral hiatus is palpated. After sterile skin preparation, a needle or intravenous catheter (18–23 gauge) is advanced at a 45° angle cephalad until a pop is felt as the needle pierces the sacrococcygeal ligament. The angle of the needle is then flattened and advanced (Figure 45–22). Aspiration for blood and CSF is performed, and, if negative, injection can proceed. Some clinicians recommend test dosing as with other epidural techniques, although many simply rely on incremental dosing with frequent aspiration. Tachycardia (if epinephrine is used) or increasing size of the T waves on electrocardiography may indicate intravascular injection. Complications are fortunately infrequent but include total spinal and intravenous injection, causing seizure or cardiac arrest. Ultrasound has also been employed in the performance of caudal blocks.

A dosage of 0.5 to 1.0 mL/kg of 0.125% to 0.25% bupivacaine (or ropivacaine), with or without epinephrine, can be used. Opioids may also be added (eg, 30–40 mcg/kg of morphine). Clonidine is often included, as well. The analgesic effects of the block may extend for hours into the postoperative period.

In adults undergoing anorectal procedures, caudal anesthesia can provide dense sacral sensory blockade with limited cephalad spread. Furthermore, the injection can be given with the patient in the prone jackknife position, which is used for surgery (Figure 45–23). A dose of 15 to 20 mL of 1.5% to 2.0% lidocaine, with or without epinephrine, is usually effective. Fentanyl, 50 to 100 mcg, may also be added. This technique should be avoided in patients with pilonidal cysts because the needle may pass through the cyst track and can potentially introduce bacteria into the caudal epidural space.

The complications of epidural, spinal, or caudal anesthetics range from mild to crippling and life-threatening (Table 45–6). Broadly, complications from neuraxial techniques are secondary to excessive physiological effects of an appropriately injected drug, injury from needle or catheter placement, and local anesthetic systemic toxicity.

A very large survey of regional anesthetics provides an indication of the relatively low incidence of serious complications from spinal and epidural anesthesia (Table 45–7). The American Society of Anesthesiologists (ASA) Closed Claims Project helps to identify the most common causes of liability claims involving regional anesthesia in the operating room setting. In a 20-year period (1980–1999), regional anesthesia accounted for 18% of all liability claims. In the majority of these claims, the injuries were judged as temporary or nondisabling (64%). Serious injuries in the remaining claims included death (13%), permanent nerve injury (10%), permanent brain damage (8%), and other permanent injuries (4%). The majority of regional anesthesia claims involved either lumbar epidural anesthesia (42%) or spinal anesthesia (34%) and tended to occur mostly in obstetric patients.

Complications Associated with Excessive Responses to Appropriately Placed Drug

A. High Neural Blockade

Exaggerated dermatomal spread of neural blockade can occur readily with either spinal or epidural anesthesia. Administration of an excessive dose, failure to reduce standard doses in selected patients (eg, the elderly, pregnant, obese, or very short), or unusual sensitivity or spread of local anesthetic may be responsible. Patients may complain of dyspnea and have numbness or weakness in the upper extremities. Nausea often precedes hypotension. Once exaggerated spread of anesthesia is recognized, patients should be reassured, oxygen supplementation may be required, and bradycardia and hypotension should be treated.

Spinal anesthesia ascending into the cervical levels causes severe hypotension, bradycardia, and respiratory insufficiency. Unconsciousness, apnea, and hypotension resulting from high levels of spinal anesthesia are referred to as a “high spinal,” or when the block extends to cranial nerves, as a “total spinal.” These conditions can also occur following attempted epidural or caudal anesthesia if there is accidental intrathecal injection (see later discussion). Apnea is more often the result of severe sustained hypotension and medullary hypoperfusion than a response to phrenic nerve paralysis from anesthesia of C3–C5 roots. Anterior spinal artery syndrome has been reported following neuraxial anesthesia, presumably due to prolonged severe hypotension together with an increase in intraspinal pressure.

Treatment of an excessively high neuraxial block involves maintaining adequate arterial oxygenation and ventilation and supporting the circulation. When respiratory insufficiency becomes evident, in addition to supplemental oxygen and assisted ventilation, intubation and mechanical ventilation may be necessary. Hypotension can be treated with intravenous vasopressors and rapid administration of intravenous fluids. Bradycardia can be treated early with atropine. Ephedrine or epinephrine can also increase heart rate and arterial blood pressure.

B. Cardiac Arrest During Spinal Anesthesia

Examination of data from the ASA Closed Claim Project identified several cases of cardiac arrest during spinal anesthesia. Because many of these cases predated the routine use of pulse oximetry, oversedation and unrecognized hypoventilation and hypoxia may have contributed. However, large prospective studies continue to report a relatively high incidence (perhaps as high as 1:1500) of cardiac arrest in patients having received a spinal anesthetic. Many of these cardiac arrests were preceded by bradycardia, and many occurred in young healthy patients. To prevent this from occurring, hypovolemia should be corrected. Prompt drug treatment of hypotension and bradycardia are recommended.

C. Urinary Retention

Local anesthetic block of S2–S4 root fibers decreases urinary bladder tone and inhibits the voiding reflex. Epidural opioids can also interfere with normal voiding.

Complications Associated with Needle or Catheter Insertion

A. Inadequate Anesthesia or Analgesia

As with other regional anesthesia techniques, neuraxial blocks are associated with a small, but measureable, failure rate that is usually inversely proportional to the clinician’s experience. Failure may still occur, even when CSF is obtained during spinal anesthesia. Movement of the needle during injection, incomplete entry of the needle opening into the subarachnoid space, subdural injection, or injection of the local anesthetic solution into a nerve root sleeve may be responsible. Causes for failed epidural blocks were discussed earlier (see “Failed Epidural Blocks”).

B. Intravascular Injection

Accidental intravascular injection of the local anesthetic for epidural and caudal anesthesia can produce very high serum drug levels, which may affect the central nervous system (seizure and unconsciousness) and the cardiovascular system (hypotension, arrhythmias, and depressed contractility). Because the dosage of medication for spinal anesthesia is relatively small, this complication is seen after epidural and caudal (but not spinal) blocks. Local anesthetic may be injected directly into a vessel through a needle or later through a catheter that has entered a blood vessel (vein). The incidence of intravascular injection can be minimized by carefully aspirating the needle (or catheter) before every injection, using a test dose, always injecting local anesthetic in incremental doses, and close observation for early signs of intravascular injection (tinnitus, lingual sensations). Advanced cardiac life support should be initiated if cardiac arrest occurs. Lipid emulsion, 20% 1.5 mL/kg bolus, should be given followed by a 0.25-mL/kg infusion. Incremental 1 mcg/kg doses of epinephrine should be administered rather than larger 10 mcg/kg doses. Should cardiac function not be restored additional lipid emulsion can be administered up to 10 mL/kg. Cardiopulmonary bypass can be used should the patient fail to respond to resuscitative efforts. A recent review of local anesthetic systemic toxicity (LAST) cases suggests that use of lipid emulsion treatment may be underutilized even when indicated.

Local anesthetics vary in their propensity to produce severe cardiac toxicity. The rank order of local anesthetic potency at producing seizures and cardiac toxicity is the same as the rank order for potency at nerve blocks. Chloroprocaine has relatively low potency and also is metabolized very rapidly; lidocaine and mepivacaine are intermediate in potency and toxicity; and levobupivacaine, ropivacaine, bupivacaine, and tetracaine are most potent and toxic.

C. Total Spinal Anesthesia

Total spinal anesthesia can occur following attempted epidural or caudal anesthesia if there is accidental intrathecal injection. Onset is usually rapid, because the amount of anesthetic required for epidural and caudal anesthesia is 5 to 10 times that required for spinal anesthesia. Careful aspiration, use of a test dose, and incremental injection techniques (remember, “every dose is a test dose”) during epidural and caudal anesthesia can help avoid this complication.

D. Subdural Injection

Because of the larger amount of local anesthetic administered, accidental subdural injection of local anesthetic during attempted epidural anesthesia is much more serious than during attempted spinal anesthesia. A subdural injection of epidural doses of local anesthetic produces a clinical presentation similar to that of high spinal anesthesia, with the exception that the onset may be delayed for 15 to 30 min and the block may be “patchy.” The spinal subdural space is a potential space between the dura and the arachnoid that extends intracranially, so that anesthetic injected into the spinal subdural space can ascend to higher levels than epidural medications. As with high spinal anesthesia, treatment is supportive and may require intubation, mechanical ventilation, and cardiovascular support. The effects generally last from one to several hours.

E. Backache

As a needle passes through skin, subcutaneous tissues, muscle, and ligaments it causes varying degrees of tissue trauma. Bruising and a localized inflammatory response with or without reflex muscle spasm may be responsible for postoperative backache. One should remember that up to 25% to 30% of patients receiving general anesthesia also complain of backache postoperatively, and a significant percentage of the general population has chronic back pain. Postoperative back soreness or ache is usually mild and self-limited, although it may last for a number of weeks. If treatment is sought, acetaminophen or NSAIDs should suffice. Although backache is usually benign, it may be an important clinical sign of much more serious complications, such as epidural hematoma and abscess (see later discussion).

F. Postdural Puncture Headache

Any breach of the dura may result in a postdural puncture headache (PDPH). This may follow a diagnostic lumbar puncture, a myelogram, a spinal anesthetic, or an epidural “wet tap” in which the epidural needle passed through the epidural space and entered the subarachnoid space. Similarly, an epidural catheter might puncture the dura at any time and result in PDPH. An epidural wet tap is usually immediately recognized as CSF pouring from the epidural needle or aspirated from an epidural catheter. However, PDPH can follow a seemingly uncomplicated epidural anesthetic and may be the result of just the tip of the needle scratching through the dura. Typically, PDPH is bilateral, frontal or retroorbital, or occipital and extends into the neck. It may be throbbing or constant and associated with photophobia and nausea. The hallmark of PDPH is its association with body position. The pain is aggravated by sitting or standing and relieved or decreased by lying down flat. The onset of headache is usually 12 to 72 h following the procedure; however, it may be seen almost immediately.

PDPH is believed to result from leakage of CSF from a dural defect and subsequent intracranial hypotension. Loss of CSF at a rate faster than it can be produced causes traction on structures supporting the brain, particularly the meninges, dura, and tentorium. Increased traction on blood vessels and cranial nerves may also contribute to the pain. Traction on the cranial nerves may occasionally cause diplopia (usually the sixth cranial nerve) and tinnitus. Subdural hematoma can occur following tear of intracerebral veins secondary to intracranial hypotension due to CSF loss. The incidence of PDPH is strongly related to needle size, needle type, and patient population. The larger the needle, the greater the likelihood that PDPH will occur. Cutting-point needles are associated with a higher incidence of PDPH than pencil-point needles of the same gauge. Factors that increase the risk of PDPH include young age, female sex, and pregnancy. The greatest risk, then, would be expected following an accidental wet tap with a large epidural needle in a young woman (perhaps as high as 20% to 50%). The lowest incidence would be expected in an elderly male using a 27-gauge pencil-point needle (<1%). Studies of obstetric patients undergoing spinal anesthesia for cesarean delivery with small-gauge pencil-point needles have shown rates as low as 3% or 4%.

Conservative treatment involves recumbent positioning, analgesics, intravenous or oral fluid administration, and caffeine. Keeping the patient supine will decrease the hydrostatic pressure driving fluid out of the dural hole and minimize the headache. Analgesic medication may range from acetaminophen to NSAIDs and opioids. Hydration and caffeine work to stimulate production of CSF. Caffeine further helps by vasoconstricting intracranial vessels as cerebral vasodilation is thought to be a response to the intracranial hypotension secondary to the CSF leak. Stool softeners and soft diet are used to minimize Valsalva straining. Headache may persist for days, despite conservative therapy.

An epidural blood patch is an effective treatment for PDPH. It involves injecting 15 to 20 mL of autologous blood into the epidural space at, or one interspace below, the level of the dural puncture. It is believed to stop further leakage of CSF by either mass effect or coagulation. The effect is usually immediate and complete, but may take some hours as CSF production slowly rebuilds intracranial pressure. Approximately 90% of patients will respond to a single blood patch, and 90% of initial nonresponders will obtain relief from a second injection. We do not recommend prophylactic blood patching after a wet tap. Not all patients will develop PDPH following a wet tap, and in any case, prophylaxis has not proven terribly effective. When evaluating patients with presumed PDPH, other sources of headache, including migraine, caffeine withdrawal, meningeal infection, and subarachnoid hemorrhage, should be considered in the differential diagnosis.

G. Neurological Injury

Perhaps no complication is more perplexing or distressing than persistent neurological deficits following an apparently routine neuraxial block. An epidural hematoma or abscess must be ruled out. Either nerve roots or spinal cord may be injured. The latter may be avoided if the neuraxial blockade is performed below the termination of the conus (L1 in adults and L3 in children). Postoperative peripheral neuropathies can be due to direct physical trauma to nerve roots. Although most resolve spontaneously, some are permanent. Any sustained paresthesia during neuraxial anesthesia/analgesia should alert the clinician to redirect the needle. Injections should be immediately stopped and the needle withdrawn, if injection is associated with pain. Direct injection into the spinal cord can cause paraplegia. Damage to the conus medullaris may cause isolated sacral nerve dysfunction. Not all neurological deficits that are reported after a regional anesthetic are the direct result of the block. Postpartum neurological deficits, including lateral femoral cutaneous neuropathy and foot drop, were recognized as complications before the era of routine epidural anesthesia/analgesia.

H. Spinal or Epidural Hematoma

Needle or catheter trauma to epidural veins often causes minor bleeding in the spinal canal, although this usually has no consequences. The incidence of spinal hematomas has been estimated to be about 1:150,000 for epidural blocks and 1:220,000 for spinal anesthetics. The vast majority of reported cases have occurred in patients with abnormal coagulation secondary to either disease or drugs. Some hematomas have occurred immediately following removal of an epidural catheter. Thus, both insertion and removal of an epidural catheter can lead to epidural hematoma formation.

Diagnosis and treatment must be prompt, if permanent neurological sequelae secondary to neuronal ischemia are to be avoided. The onset of symptoms is typically more sudden than with epidural abscess. Symptoms include sharp back and leg pain with a motor weakness or sphincter dysfunction, or both. When hematoma is suspected, imaging (magnetic resonance [MR] or computed tomography [CT]) and neurosurgical consultation must be obtained immediately. In many cases, good neurological recovery has occurred in patients who have undergone prompt surgical decompression.

Neuraxial anesthesia should be avoided in patients with coagulopathy, significant thrombocytopenia, platelet dysfunction, or those who have received fibrinolytic or thrombolytic therapy. Practice guidelines should be reviewed when considering neuraxial anesthesia in such patients, and the risk versus benefit of these techniques should be weighed and delineated in the informed consent process.

I. Meningitis and Arachnoiditis

Infection of the subarachnoid space can follow neuraxial blocks as the result of contamination of the equipment or injected solutions, or as a result of organisms tracked in from the skin. Indwelling catheters may become colonized with skin organisms.

Strict sterile technique should be employed. Careful attention is particularly warranted in the labor room, where family members are often curious to see what is being done to mitigate the parturient’s pain. If hospital policy permits their presence during epidural placement, such individuals should be advised to avoid contaminating the tray. Family members should also wear a mask to prevent contamination of the epidural tray with oral flora.

J. Epidural Abscess

Spinal epidural abscess (EA) is a rare but potentially devastating complication of neuraxial anesthesia. The reported incidence varies widely, from 1:6500 to 1:500,000 epidurals. Most reported anesthesia-related cases involve epidural catheters. In one reported series, there was a mean of 5 days from catheter insertion to the development of symptoms, although presentation can be delayed for weeks.

There are four classic clinical stages of EA, although progression and time course can vary. Initially, symptoms include back pain that is intensified by percussion over the spine. Second, nerve root or radicular pain develops. The third stage is marked by motor or sensory deficits, or sphincter dysfunction. Paraplegia or paralysis marks the fourth stage. Ideally, the diagnosis is made in the early stages. Prognosis has consistently been shown to correlate to the degree of neurological dysfunction at the time the diagnosis is made. Back pain and fever after epidural anesthesia should alert the clinician to consider EA. Radicular pain or neurological deficit heightens the urgency to investigate. Once EA is suspected, the catheter should be removed (if still present) and the tip cultured. The injection site is examined for evidence of infection; if pus is expressed, it is sent for culture. Blood cultures should be obtained. If suspicion is high and cultures have been obtained, anti-Staphylococcus coverage can be instituted, as the most common organisms causing EA are Staphylococcus aureus and Staphylococcus epidermidis. MR or CT imaging should be performed to confirm or rule out the diagnosis. We recommend prompt consultation with specialists in neurosurgical and infectious disease. In addition to antibiotics, treatment of EA usually involves decompression (laminectomy), although percutaneous drainage with fluoroscopic or CT guidance has been used. Suggested strategies for guarding against the occurrence of EA include (1) minimizing catheter manipulations and maintaining a closed system when possible; (2) using a micropore (0.22-μm) bacterial filter; and (3) removing an epidural catheter or at least changing the catheter, filter, and solution after a defined time interval (eg, some clinicians replace or remove all epidurals after 4 days).

K. Sheering of an Epidural Catheter

There is a risk of neuraxial catheters sheering and breaking off inside of tissues if they are withdrawn through the needle. If a catheter must be withdrawn while the needle remains in situ, both must be carefully withdrawn together. If a catheter breaks off within the epidural space, many experts suggest leaving it and observing the patient. If, however, the breakage occurs in superficial tissues, the catheter should be surgically removed.

Complications Associated with Drug Toxicity

A. Local Anesthetic Systemic Toxicity

Absorption of excessive amounts of local anesthetics can produce toxic blood levels (see “Intravascular Injection”). Excessive absorption from epidural or caudal blocks is rare when appropriate doses of local anesthetic are used.

B. Transient Neurological Symptoms

First described in 1993, transient neurological symptoms (TNS), also referred to as transient radicular irritation, are characterized by back pain radiating to the legs without sensory or motor deficits, occurring after the resolution of spinal anesthesia and resolving spontaneously within several days. It is most commonly associated with hyperbaric lidocaine (incidence up to 12%), but has also been reported with tetracaine (2%), bupivacaine (1%), mepivacaine, prilocaine, procaine, and subarachnoid ropivacaine. There are also case reports of TNS following epidural anesthesia. The incidence of this syndrome is greatest among outpatients, particularly males undergoing surgery in the lithotomy position, and least among inpatients undergoing surgery in positions other than lithotomy. The pathogenesis of TNS is believed to represent concentration-dependent neurotoxicity of local anesthetics.