Chapter 47. Neuraxial Anesthesia

1. A systematic and rational approach based on a thorough 3-dimensional understanding of anatomy should be used when accessing the subarachnoid or epidural space.

2. Anesthetic doses, agents, and combinations of agents should be individualized to optimize neuraxial blockade for a given clinical setting.

3. Hypotension and bradycardia associated with neuraxial anesthesia should be identified early and treated aggressively to minimize development of cardiovascular collapse and poor outcome.

4. Our understanding of potential neurotoxicity and the nature of transient neurologic symptoms (TNS) are continuing to evolve. However, there is growing consensus that TNS may not represent direct neural toxicity.

5. Evaluating the appropriateness of neuraxial procedures in patients receiving anticoagulant and antiplatelet medications is a challenge. Clinicians should be familiar with the recommendations presented by the American Society of Regional Anesthesia and Pain Medicine in the consensus statement addressing these issues.

6. When suspicion of spinal hematoma or abscess is credible, definitive diagnosis with appropriate imaging and prompt surgical decompression within 4 to 8 hours of onset of neurologic symptoms is crucial to improve chances of recovery of function.

7. Developing an understanding of the nature of combined spinal–epidural anesthesia and facility with its techniques can expand a clinician's armamentarium to provide neuraxial anesthesia and optimize patient care.

With more than 100 years of use, neuraxial anesthesia has enjoyed much success and endured controversy. With the stage set by the developments of the hollow needle and syringe, the discovery of neuraxial anesthesia began in 1885 when Corning was experimenting with effects of cocaine on the spinal nerves of dogs. Bier brought spinal anesthesia into clinical use for surgery in 1898, but only after self-experimentation and a personal experience with a well-described postmenigeal puncture headache. Epidural anesthesia gained widespread attention in the setting of labor analgesia, maintaining the medical community's interest in neuraxial blockade despite rapid advancements in techniques for general anesthesia in the 1940s and 1950s. More recently, the introduction of continuous catheter techniques, combined spinal–epidural anesthesia, and various neuraxial anesthetic adjuvants has presented further opportunities to provide our patients the many benefits of neuraxial blockade.

Neural Structures

A thorough appreciation for the anatomy of spinal structures is necessary for appropriate technique, patient selection, and management of neuraxial anesthesia. The spinal structures are identified by cervical, thoracic, lumbar, sacral, and caudal regions (Fig. 47-1). The spinal cord begins at the base of the brainstem and continues caudad terminating as the conus medullaris, typically at the L1-L2 level in the adult, although the terminus rarely may be as high as T12 or as low as L4. The spinal cord has cervical and lumbar enlargements to accommodate the increased neuronal supply to the limbs. Rootlets emerge from the dorsal and ventral surface of the spinal cord and converge to form the respective ventral and dorsal roots of the spinal nerves at each spinal level (Fig. 47-2).

Cerebrospinal Fluid

As with the brain, the spinal cord and portions of the spinal nerves are bathed in cerebrospinal fluid (CSF), which provides protection for these structures and participates in maintaining homeostasis. The CSF is secreted by the choroid plexuses located on the roofs of the lateral, third, and fourth ventricles, accounting for 40% to 60% of production. Water derived from glucose metabolism is also a major CSF source. CSF is reabsorbed mainly through arachnoid villi, which project into the parasagittal venous sinuses, and to a much lesser extent into the epidural veins accompanying spinal nerves. CSF typically has a density range of 1.00028 to 1.00100 g/mL.1 CSF volume and its clinical significance are discussed in the Clinical Pharmacology section.

Meninges

As with the brain, the spinal cord, nerve roots, and CSF are enveloped by 3 membranes: the pia mater, arachnoid mater, and the dura mater (Figs. 47-2 and 47-3). The innermost pia mater is intimately applied to the surface of the cord and nerve roots. It is a vascular and permeable membrane. The arachnoid mater approximates the outermost membrane, the dura mater. Between the pia and the arachnoid is the subarachnoid space, home to the CSF, arachnoid trabecular network, and dentate ligaments. The subarachnoid space extends laterally with the dura as nerve root "sleeves," with the accompanying CSF as far as the dorsal root ganglion. The spinal dura mater extends from the level of the foramen magnum, where it is contiguous with cranial dura, to the level of S2 in most adults. Here the subarachnoid space terminates as the dura fuses with the filum terminale (pial extension) and then with the coccygeal periosteum. Between the dura mater and arachnoid mater resides the subdural space, which contains little more than a small amount of serous fluid. This space is often implicated, though not proved, as a link in complications such as a total spinal after an intended epidural administration of local anesthetic and failed spinal anesthetics.

Vertebrae

The neural structures of the spinal cord are protected by the bony vertebral column, which is composed of 7 cervical, 12 thoracic, 5 lumbar, 5 sacral, and 4 coccygeal vertebrae (Fig. 47-1). The vertebrae of the sacral and coccygeal regions are fused to form the sacrum and coccyx, respectively. Cervical and lumbar lordotic curves and thoracic kyphotic curves of the normal spine allow for distribution of mechanical forces and spinal movements. A typical vertebra consists of a pillar-like vertebral body joined by the pedicles to the posterior elements, namely the laminae, superior and inferior articular processes, transverse processes, and the spinous process. The vertebrae have regional characteristics that are relevant to the techniques used to access the neuraxis in clinical situations. The figures depict relevant features of representative thoracic and lumbar vertebrae (Fig. 47-4).

Intervertebral Disks and Ligaments

The endplates of the vertebral bodies are joined to one another by the intervertebral disks (Fig. 47-2), consisting of the outer annulus fibrosus and the inner nucleus pulposus. Further structural support comes from the anterior and posterior longitudinal ligaments running along the ventral and dorsal aspects of the vertebral bodies respectively. Along with the supraspinous ligament traversing superficially, the deeper interspinous ligament joins the spinous processes to one another (Fig. 47-3). Adjacent laminae are interconnected by the ligamentum flavum, together forming the posterior wall of the spinal canal. The ligamentum flavum forms embryologically from 2 (right and left) separate neural crest structures. These ligamenta meet in the midline at an acute angle, but have an inconsistent degree of fusion. Incomplete fusion results in sagittal gaps, which are variable not only between individuals, but also among spinal levels within a given patient. Cervical and upper thoracic levels have a higher rate of failed fusion compared with levels below T3-T4 (as high as 50%-70%) and are thus perhaps more difficult for epidural access using the loss or resistance technique.2

Epidural Space

The epidural space (also referred to as extradural and peridural space) lies between the dura and the borders of the spinal canal (Fig. 47-3). Anteriorly, this border is formed by the posterior longitudinal ligament; posteriorly, by the vertebral lamina and adjoining ligamentum flavum (Fig. 47-3). The spinal epidural space runs from the level of the foramen magnum to the sacral hiatus, which is bound by the sacrococcygeal ligament. The lateral borders of the epidural space are partially delineated by the vertebral pedicles, but this space extends laterally through the intervertebral foramina to communicate with the paravertebral spaces on each side. The epidural space is somewhat compartmentalized by the sections of the dura abutting the ligamentum flavum, vertebral lamina, and other borders of the vertebral canal (Fig. 47-3B). However, these compartments are joined by a "potential space" that is opened by injection of fluid or air, thus connecting the compartments and revealing a more continuous communication.

The contents of the epidural space include epidural fat, venous plexus, and segmental arteries. The epidural fat is largely located in the posterior and lateral aspects of the epidural space. This fat can be a site for sequestration of epidurally administered local anesthetics and opioids as a function of the lipid solubility of the drug.3 The plexus of epidural veins (Batson's plexus) is principally within the anterior and lateral portions of the epidural space, with rare presence in the posterior aspect. These veins communicate with the azygous system, which can become engorged in the setting of increased intra-abdominal or intrapelvic pressure.

Spinal Needles

Although many designs have been introduced throughout the history of spinal anesthesia, the needles commonly used today can be generally classified as either cutting needles or pencil-point needles (Fig. 47-5). The Quincke-style needle, one of the most frequently used cutting needles, has a medium-length bevel, coming to a sharp point. In contrast, pencil-point needles, such as the Whitacre, Sprotte, and Gertie-Marx, have a rounded, noncutting tip. The latter 3 needles have openings on the side rather than at the tip. Efforts to minimize the incidence of postmeningeal puncture headache have driven the advancement of needle design. More recent designs include tapered needles and return of stylet-point needles, such as the Ballpen needle (Rusch, France). Needle gauge usually ranges from 27 to 18, with the thinner needles requiring an introducer needle to prevent spinal needle deflection.

Epidural Needles and Catheters

Needles for accessing the epidural space need to have a large enough diameter (eg, 16-20 gauge) to permit catheter insertion and facilitate loss of resistance techniques to identify the epidural space.

Currently, the most commonly used needle is the Tuohy needle (Fig. 47-5D). The design of this needle has undergone many modifications, but is characterized by a curved, or Huber, tip. This curved tip decreases coring of tissue during insertion, but has been used chiefly in an effort to provide directional control of catheter insertion. Epidural needles have a tight-fitting stylet, which prevents coring of the skin with subsequent transfer to the epidural space where it can grow and cause mass effects. Larger-gauge needles are less likely to be deflected by firm ligaments and osseous structures and may provide a more reliable "loss-of-resistance" for identification of the epidural space.4

Commercially available catheters for cannulation of the epidural space are composed of polyurethane, nylon, or silicone-based material. Catheters with wire-wound reinforcement are designed to prevent kinking of the lumen and to increase durability, thus making them useful for postoperative analgesic infusions. Several variations of catheters have been brought to the market, with options for multi-orifice tips, metal stylets, and varying stiffness. Although the many available needles and catheters have various purported advantages, definitive, consistent data for superiority for any particular product are limited. Therefore, selection among these is primarily driven by subjective preference or availability.

The clinical success of neuraxial anesthesia begins with proper patient selection and creation of an environment that is conducive to regional anesthetic techniques. Principally, the surgical procedure should be able to be performed with a sensory and motor block level that is tolerable and safe for the patient under the appropriate positioning, monitoring, and supplemental medication. Lower-extremity, pelvic, and lower-abdominal procedures are often appropriate for spinal or epidural anesthesia. Procedures in the perineum, or involving the foot, usually are more suited to spinal anesthesia because of the potential for sacral nerve sparing with epidural anesthesia. Upper abdominal procedures usually require sensory levels that are not comfortably attainable with regional techniques alone and would require "supplementation" with light general anesthesia.

It is imperative that the patient be able to cooperate and tolerate the experience, from placement of the block, to block resolution. This can be accomplished through proper preoperative evaluation of the patient's maturity and affect and appropriate informed consent discussions regarding risks, benefits, alternatives, and patient expectations in light of the requirements of the procedure.

Absolute contraindications to neuraxial anesthesia are few in number, but include patient refusal, significant hypovolemia, infection at the site of needle entrance, increased intracranial pressure, and significant coagulopathy. Relative contraindications require analysis of the risks and benefits for a given patient and situation. Unfortunately, this often needs to be done in light of the current medicolegal environment. Commonly encountered conditions that might be considered relative contraindications are minor coagulopathy, sepsis/bacteremia, and preexisting neurologic conditions, such as peripheral neuropathy, multiple sclerosis, and other demyelinating processes. Although definitive data for the relative safety of neuraxial anesthesia in patients with central or peripheral neurologic disorders are lacking, the ASRA Practice Advisory in Neurologic Complications in Regional Anesthesia and Pain Medicine suggests that the risk of new neurologic deficit or exacerbation of a present condition appears to be extremely rare in these conditions.5

Patient Positioning and Preparation

In addition to confirming patient consent and understanding, one should also verify the immediate availability of equipment and medications to manage airway compromise and hemodynamic perturbations. The patient should have proper monitoring and supplemental oxygen during neuraxial anesthesia. Most patients prefer to have sedation administered before beginning neuraxial anesthesia, and this is typically achieved with judicious use of agents such as midazolam and fentanyl, maintaining the patient's ability to report paresthesia. Performance of neuraxial anesthesia on anesthetized or heavily sedated patients is considered routine in pediatric anesthesia, but remains controversial in adult care. Although wakefulness is suggested for adults as a means of decreasing the risk of neurologic injury,5 some clinical situations may present with extenuating circumstances that justify deep sedation while performing neuraxial procedures after a careful consideration of the risks and benefits of the practice.

The lateral decubitus position (Fig. 47-6) is most often used for spinal anesthesia. The patient should have hips and knees flexed with head approximating the knees in an effort to relax the lumbar lordotic curve and accentuate the interlaminar aperture.

Alternatively, neuraxial access can also be obtained with the patient in the sitting position (Fig. 47-7), bending forward to relax the lordotic curve of the lumbar spine. Advantages of this position are that it facilitates the identification of the midline in obese or anatomically distorted patients and augments CSF pressure at the needle entrance site, enhancing CSF flow through the needle when performing lumbar puncture. It is also used to accomplish "saddle block" anesthesia for procedures limited to the perineum by administering hyperbaric spinal preparations (see section on Baricity) to the intrathecal space to promote caudal migration of the anesthetic agent. If the patient is going to be in the prone position for procedures such as perianal surgery, a hypobaric technique can be used. This is accomplished by positioning the patient for surgery prone with hips flexed and lumbar spine relaxed (jackknife) (Fig. 47-8), then accessing the subarachnoid space, using a syringe to aspirate for confirmation of CSF flow. Once a hypobaric solution is administered, it is recommended that the patient remain in a flat or head-down position for at least 1 hour to prevent unintended rostral migration of the anesthetic.

Accessing the Subarachnoid Space

It is recommended that all procedures for neuraxial anesthesia begin with foundational elements of safety, including protective masks for all participants (US Centers for Disease Control and Prevention [CDC] guideline), marking of the target site (if lateralized), setup of procedural tray and equipment (including labeling of all medications and solutions), and a final procedural pause immediately before starting the procedure to review critical information (correct patient, correct procedure, lack of contraindications, appropriate equipment available, etc) Several techniques have been described for inserting a needle percutaneously into the subarachnoid space, each having specific advantages in different clinical situations. By developing skill with multiple techniques, the practitioner can be prepared to adapt to the clinical challenges presented. Regardless of the position and technique to be used, the procedure should begin with identifying the midline and palpating the iliac crests. The intercrestal line (Fig. 47-6) is classically described as crossing the level of the fourth lumbar vertebra, or the L3-L4 or L4-L5 interspace, which are the preferred levels of entry, due to increased risk of spinal cord injury when entering at higher vertebral levels. The most commonly used and easily mastered technique for needle insertion is the midline approach. After identification of the desired interspace, a skin wheal is raised over the interspace, and appropriate infiltration of local anesthetic is fanned along the midline. It is useful to firmly place 2 fingers of the nondominant hand parallel to the axis, straddling the lateral borders of the interspace (Fig. 47-9). When using small-diameter spinal needles, an introducer needle is first inserted in the sagittal plane between the spinous processes, initially parallel to the plane of the surface of the back. The introducer is then stabilized by grasping the hub with the digits of the nondominant hand as the spinal needle is passed through the bore of the introducer and advanced in the sagittal plane. The path of travel for the needle in the midline approach is through skin, subcutaneous tissue, supraspinous ligament, interspinous ligament, ligamentum flavum, and, finally, through the dura and arachnoid mater into the subarachnoid space. As the needle tip passes beyond the ligamentum flavum and through the dura, a change in resistance, and possibly a "pop," can be appreciated. At this point, the stylet is withdrawn from the spinal needle, and the hub is inspected for free flow of CSF. If bone is contacted before entering the thecal sac, the practitioner should reassess the angle of projection and ensure that it is within the sagittal plane. If no correction toward the midline is needed, the most fruitful adjustment is usually an attempt to gradually "walk" the needle tip in a cephalad direction until the subarachnoid space is encountered. If progress in the cephalad direction leads to contacting bone at a more shallow depth, then a caudad direction should be attempted because it is likely that you have contacted the more superior lamina. If free flow of CSF is in question, rotating the bevel or side hole of the needle through the 4 quadrants can help ensure proper placement of the needle opening. In addition, because the holes of the pencil-point needles are located proximal to the tip, it is possible for the tip to be in the subarachnoid space but not the needle hole. Consequently, advancing a pencil-point needle a 1 or 2 mm may result in free flow of CSF.

Once satisfied with needle placement, the dosing syringe (typically a Luer slip style), which was previously prepared with the intended agent(s), is firmly attached to the hub of the spinal needle. Aspiration of CSF into the syringe is final confirmation of subarachnoid placement; the agent is slowly injected while keeping the needle tip immobile. Once the injection is complete, the patient should be placed in the intended position to impact the spread of the agent or into the appropriate position for surgery. The advantages of the midline approach include simplicity and the stability provided by the relatively avascular sagittal plane through the interspinous ligament.

The paramedian technique (Fig. 47-10), which is routinely used by some, should be mastered to the same level as the midline technique because of its advantages in common clinical situations. When the midline approach is difficult, the paramedian approach may still allow access to the subarachnoid space in patients who are unable to maintain ideal positioning, are obese or pregnant, or who have heavily calcified interspinous ligaments (common in elderly patients), compression fracture, or scoliosis. This advantage is a consequence of the larger interlaminar aperture presented via the path of the paramedian approach, especially when lumbar lordosis is retained. Although variations have been described, in the lumbar region this approach begins with identification of the superior aspect of the caudad spinous process of the interspace to be accessed. The percutaneous entry point is identified 1 cm lateral to this superior tip of the spinous process. After raising a skin wheal and appropriate infiltration with local anesthetic, the introducer and/or spinal needle are advanced through the anesthetized region, directed approximately 10 to 15 degrees toward midline and at a slightly cephalad angle. This path will traverse skin, subcutaneous tissue, paraspinous muscle, ligamentum flavum, dura, and arachnoid mater to reach the subarachnoid space. If bone is encountered prior to dural puncture, it is most likely the caudad lamina, in which case the needle should be "walked" cephalad, then medial, until the thecal sac is breached.

The Taylor approach (Fig. 47-11) is a variation of the paramedian technique; it capitalizes on the breadth of the L5-S1 interspace (typically the largest interspace) and the palpable and relatively constant location of the posterior superior iliac spines (PSISs). Once a PSIS is identified, a skin wheal is raised 1 cm medial and caudad to its inferior border. After appropriate infiltration of local anesthetic, the spinal needle (typically at least 22 gauge) is directed approximately 45 degrees medial and cephalad toward the L5-S1 interspace. Again, if bone is contacted, the correction needed is a greater cephalad angle to "walk off" of the superior aspect of the sacrum.

Accessing the Epidural Space

Although the pathway and approach to the epidural space are the same for the subarachnoid space, the technical skills to correctly and reliably access and identify the epidural space require more experience to develop than those for dural puncture.6 Again, both midline and paramedian approaches are used for needle placement, and the patient can be placed in either the sitting or lateral position. In contrast to spinal anesthesia with increasing risk of cord injury above the L2 level, the epidural space theoretically can be accessed at any vertebral level (though there remains the risk of spinal cord injury above L2 if the clinician fails to correctly identify the epidural space or control the advancing needle). Because of the differences in the projecting angles of the spinous processes in the lumbar and midthoracic regions, an angle of approach closer to the plane of the surface of the back is used when performing thoracic epidural placement.

With the midline approach, the desired interspace is identified by palpation as described previously, and a skin wheal is raised with local anesthetic at the site of intended needle entrance, which should be in the sagittal plane at the midpoint between the spinous processes. The needle is inserted through the skin, subcutaneous tissue, and supraspinous ligament and into the interspinous ligament. With the blunter tip of the larger-diameter needles used for epidural anesthesia, passage into the interspinous ligament is typically met with a "crunchy" sensation. Patients who are not sedated might either hear or feel this "crunching" and should be reassured if report of this is elicited. If in the sagittal plane, further advancement of the needle should place the tip in the ligamentum flavum, which provides increased resistance to needle advance.

Alternatively, if the paramedian approach is used, as may be required in the thoracic region due to the acutely angled spinous processes, the interspinous ligament is not typically encountered. If attempting access in the thoracic region, the appropriate needle entrance site is 1 cm lateral to the inferior aspect of the superior spinous process of the interspace desired. The needle is advanced through an anesthetized skin wheal, subcutaneous tissues, paraspinous muscle, and to the lamina of the inferior vertebra, providing a reference depth. The needle tip is walked medially until the base of the spinous process is encountered, which will be noted to have a shallower depth. The needle tip is then marched superiorly until it "walks off" of the lamina, and it should encounter resistance as it meets the ligamentum flavum.

With either approach, correct passage of the needle tip through the ligamentum flavum and into the epidural space without breaching the dura can be accomplished by using several different methods for identifying the epidural space. The most commonly used method is the loss of resistance (LOR) technique using saline or air, although saline may be preferred, given the risk of headache if air is unintentionally injected into the subarachnoid space. A low-resistance syringe made of glass or plastic is filled with 1 to 2 mL of sterile saline, as well as a small (0.2 mL) air bubble (Fig. 47-12). This syringe is attached to the epidural needle once the tip has been placed in the ligamentum flavum. The degree of resistance to injection of the saline is assessed by determining whether the bubble can be compressed without allowing injection of fluid. If little resistance is encountered, the needle tip is either not yet within the ligamentum flavum, or perhaps has passed through to the epidural space. If uncertain, careful advancement of the needle by a few millimeters could be considered in an attempt to reach the ligament, but one might wish to withdraw and reattempt placement. Once this ligament is identified, the needle is advanced slowly, with constant pressure on the syringe plunger, until the resistance to injection is reduced. A dramatic loss of resistance is reassuring for passage of the needle tip just beyond the ligamentum flavum and into the epidural space.

An alternative technique is the "hanging drop" of Gutierrez. This relies on negative pressure in the epidural space to aspirate a visible drop of fluid "hanging" from the hub of the needle and into the lumen of the epidural space itself (Fig. 47-13). This negative pressure is encountered just as the tip of the needle enters the epidural space and is much more reliable in the thoracic and cervical regions. One theory for the origin of this negative pressure is that it is created by the tip of the needle tenting the dura, thus increasing risk of dural puncture. This technique is best used in the thoracic region with a sitting patient. It may be helpful in the setting of combined spinal–epidural anesthesia because of minimal saline entering the epidural space and less chance of a "false return" being interpreted as CSF.7

Newer techniques for identifying the epidural space are being described. The use of electrical nerve stimulation via a conducting catheter allows for confirmation of placement of the catheter within the epidural space by confirming appropriate stimulation threshold.8 Additionally, the myotomal distribution of the stimulation can provide information regarding the spinal level of the catheter tip. Lechner et al9 describe an apparatus for use in epidural access that produces both audible and visual indication of the pressure encountered at the tip of an attached epidural needle. Using this device allows a 2-handed (which facilitates control) advancement of the epidural needle while awaiting the drop in pressure that is characteristic of entering the epidural space. Further testing and study are needed to determine the appropriate role for these new methods.

As ultrasound-guided regional anesthesia continues to gain popularity, some have attempted to apply these tools to the performance of neuraxial anesthesia. At the time of this writing, the use of ultrasound for neuraxial anesthesia is primarily described in pediatric and obstetric anesthesia for initial evaluation of anatomic landmarks and estimation of depth of the epidural space. In these populations, it has been demonstrated that the use of ultrasound can result in fewer passes at fewer interspaces, but there is no support for the blocks being more successful or safer.10 Further studies are under way that will help determine the clinical utility of and impact of ultrasound imaging on neuraxial anesthesia.

Distribution, Uptake, and Elimination

Before reaching the site of action, anesthetic agents injected into the subarachnoid space undergo dilution within the CSF via mass effect. It has long been postulated that CSF circulation played a role in distribution of anesthetic agents; however, this has not been reliably demonstrated. In fact, is now known that CSF motion is primarily a rostral-caudad oscillation with no net movement,11 rather than a circuitous flow, as has been traditionally described. The uptake of anesthetic agents by nerve roots depends on the concentration of agent at a given point, lipid content of the tissue, and hydrophobicity of the local anesthetic. Drug elimination from the subarachnoid space and spinal cord is determined by regional blood flow (primarily spinal cord and dura mater) and vascular reabsorption, not metabolism of the anesthetic within the intrathecal space. Drugs administered into the epidural space must diffuse across the spinal meninges (primarily the arachnoid mater which is the primary permeability barrier) to reach the CSF and neural tissues, where it produces conduction blockade. The rate and extent to which any drug distributes from the epidural space to the subarachnoid space depends on how much drug is taken up via competing pathways; in particular, clearance via the capillary network in the dura mater (particularly for hydrophilic drugs) or uptake into the epidural fat (particularly for hydrophobic drugs).

Factors Affecting Block Height

Spinal Anesthesia

Patient Characteristics

The ability to affect and predict sensory block height is important for producing reliable and manageable spinal anesthesia. Many parameters have been studied as potential determinants of block height (Table 47-1). Current data suggest that the most significant physiologic parameter determining block height is CSF volume. Unfortunately, easily measurable physical features, such as height, weight, sex, and age, do not correlate well with CSF volume.12 However, it has been observed that obese individuals have lower CSF volumes than those who are not obese. Although some studies show a slight influence of age and height on peak block levels,13,14 these are weak correlations with inadequate predictive power to influence clinical practice.

Baricity

Baricity (the ratio of the specific gravity of the local anesthetic solution to the specific gravity of CSF), drug dose (mass), and patient position relative to baricity of the agent are the most important factors influencing block height. Hyperbaric solutions will migrate within the CSF in the direction of gravity. Conversely, hypobaric solutions will rise opposite gravitational force in the CSF column.

Solutions with a density greater than 1.0015 g/mL (3 standard deviations above mean human CSF density) can be expected to reliably behave as a hyperbaric solution (Table 47-2). Hyperbaric solutions are typically prepared by adding dextrose to the solution. However, it should be noted that room temperature 2% 2-chloroprocaine may behave in a hyperbaric fashion. Once injected into the subarachnoid space, hyperbaric solutions distribute to the dependent portions of the spinal column. Thus, for a patient in the sitting position, the agent sinks to the sacral regions to produce a "saddle block." In the supine patient, hyperbaric agents injected at or above the apex of the lumbar lordotic curvature will migrate to the thoracic kyphosis, hence the tendency for a peak block of T4-6 in most supine patients. Manipulation of patient position after hyperbaric subarachnoid injection can enhance either migration of the agent to the thoracic region (head down position) or to the sacral region (head up or sitting position).

Isobaric solutions have a density that is clinically equivalent to that of CSF. Thus, once injected into the subarachnoid space, the agent neither sinks nor floats. This can be advantageous because patient position does not affect block height. In such a scenario, peak block height is more determined by dose of anesthetic agent (mass effect), but individual CSF density and volume contribute.19 It should be noted that it may be difficult to achieve block heights above the midthoracic region because of the tendency of isobaric solutions to remain in the lumbar region after injection due to the lack of sufficient forces promoting mixing and migration.

Hypobaric solutions have a distinct advantage in perianal surgery or other procedures that require a head down or jackknife position or for lateralized procedures when the patient cannot lie on the operative side (eg, hip fracture). Hypobaric solutions are produced by diluting anesthetic agents in distilled water to produce a density less than 0.9990 g/mL. For perirectal and perineal surgery that will be performed in the prone jackknife position, the patient is positioned prone with the head down, and an intrathecal injection is performed with a hypobaric local anesthetic solution, which will rise within the CSF to the sacral region. As mentioned previously (see Patient Positioning and Preparation), the patient must be recovered in the head-down or flat position for at least 60 minutes from time of injection to prevent unintentional rise of the block.16

Other Factors

The volume and concentration of isobaric spinal local anesthetics have been shown not to affect block height when using a constant dose.20 In contrast, when using hyperbaric preparations, baricity seems to overcome alterations in dose and volume, showing similar block heights with a range of doses if concentration is held constant. This highlights the propensity for hyperbaric preparations to settle to the dependent portions of the spinal canal.

Injection site can influence the nature of spinal blockade. If using hyperbaric solutions, injection low in the lumbar region can cause failure to achieve expected block height and cephalad block density as a result of sacral pooling and absence of agent above the lordotic curve. However, if injection is made at or above L2-3, then block height and success are more reliable, though it must be noted that risk of spinal cord injury theoretically increases above L3-4. For isobaric solutions, block height can be reduced by as much as 2 dermatomes per interspace when comparing administrations at L2-3, L3-4, and L4-5.14 The manner of injection is thought to have little influence on block characteristics. Investigations into the effect of barbotage show no impact on peak block height, although it does seem to reduce the time from injection to achievement of peak block height.21

Epidural Anesthesia

Most procedures that are amenable to spinal anesthesia can also be served by an epidural anesthetic that produces a similar block height (Table 47-3). However, unlike spinal anesthesia, lumbar epidural anesthesia is likely to spare or delay onset in sacral dermatomes, which can make its use in perineal procedures and surgeries involving the foot or lower leg problematic. Thus, when planning an epidural anesthetic, the choice of local anesthetic, dose, and injection site must be appropriate for the given procedure and the clinical context.

In contrast to spinal anesthesia, epidural anesthesia produces a segmental block that is closely related to physical spread of local anesthetic within the epidural space, which in turn is related to the site of local anesthetic injection. For example, Visser et al22 administered 3 mL of 2% lidocaine into the low, middle, and high thoracic epidural space and found that spread was primarily caudad from the high thoracic injection, primarily cephalad from the low thoracic injection, and equally distributed cephalad and caudad after the midthoracic injection. Similar results were reported by Masataka et al, who additionally measured spread after lumbar injection and found that spread was primarily cephalad from the injection site. This finding helps to explain the sacral sparing noted previously. These injection site–related differences in spread within the epidural space are presumably related to differences in epidural compliance.

Clinically, the best injection site is that which falls in the middle of the range of dermatomes targeted. Thus, for an incision extending from T9 to L1, a T11-T12 injection site would be ideal, although low to midthoracic dermatomes can be reached from lumbar injection sites if a sufficient local anesthetic dose is used.

In general, one can think of administering 2 mL of local anesthetic solution for each additional dermatome to be blocked. When administering anesthetic to the thoracic region, the dose is typically reduced by 30% to 50% as compared with the lumbar region to accommodate the greater spread of solution in the thoracic epidural space and thereby avoid unwanted cephalad spread. Greater spread is thought to be caused by the decreased volume and compliance of the thoracic epidural space.

When considering the ranges that are typically used for surgical anesthesia, concentration of the anesthetic solution has little effect on extent of the epidural block, but increased block density is observed with higher concentrations. Increasing the volume of solution injected for a given total dose will have a minor, and nonlinear, effect on cephalad spread (approximately 4 dermatomes when increasing from 10 to 30 mL), but will compromise density and quality of the block. Patient position also has no demonstrable effect on lateralization of the block or cephalad spread, but is often incorrectly used clinically.

The effect of patient age on epidural block height and dose requirement seems to be only clinically relevant in extremes of age, but some studies suggest as much as a 40% reduction in dose requirement when comparing 60- to 79-year-old patients with 20- to 39-year-old patients.24 This is attributed to the increased likelihood of spinal canal stenosis and decreased epidural volume, resulting in greater spread in the older population. Similarly, the influence of height and weight on the spread of epidural block is small25 and usually not clinically relevant unless considering the extremes of the spectrum.

The effect of pregnancy on the spread of epidural blockade is unclear. Several studies suggest greater spread and density of epidural block per milligram of local anesthetic during pregnancy. The mechanism responsible for this observation is unclear. Some early reports suggest that the gravid uterus compresses pelvic and abdominal veins, forcing blood into the epidural venous plexus and decreasing the volume of the epidural space. However, the fact that the increase is seen in early pregnancy before the uterus is significantly enlarged suggests this may not be the case.26 Interestingly, local anesthetic block of peripheral nerves is also enhanced beginning in early pregnancy,27 suggesting that sex steroid potentiation of local anesthetic action may be responsible for exaggerated epidural and spinal block in pregnancy.

In summary, patient physical characteristics seem to have little if any predictable clinical effect on epidural block spreads over the relatively narrow range of human sizes and ages typically encountered clinically. However, clinical judgment may allow for reducing the initial epidural dose for a patient who is very short, very old, and morbidly obese. Conversely, more aggressive dosing can be used for the patient who is very young, tall, and thin. Comorbidities and other clinical challenges may help guide the appropriate dosing regimen (ie, will it be easier/safer to deal with a failed block [inadequate spread and/or duration] or to treat the complications and side effects resulting from a block that is higher or longer than intended).

Local Anesthetics

Local anesthetic agents are discussed in Chapter 45 in detail, but some clinical points relating to neuraxial anesthesia are addressed here. The choice of neuraxial agents is determined by the nature and estimated duration of the surgical procedure, as well as postoperative issues such as planned discharge. Tables 47-4 and 47-5 list representative durations for doses of spinal and epidural agents. For spinal anesthesia, increasing the dose of agent administered prolongs the block, and unless counteracted by positioning and baricity, it may also increase peak block height. Longer-acting local anesthetics, such as bupivacaine, levobupivacaine, tetracaine, and ropivacaine, are typically chosen for longer procedures (>120 minutes). Lidocaine, mepivacaine, and prilocaine are considered to be of intermediate duration (60-120 minutes). Short-acting local anesthetics include procaine and 2-chloroprocaine, which are used for appropriately brief procedures (<60 minutes), particularly in outpatients.

Long-Duration Agents

Bupivacaine and levobupivacaine have the benefit of a very low incidence of transient neurologic symptoms29 (TNS; see section on Neurologic Complications); however, the variability in time to complete block resolution and achievement of discharge criteria provides a challenge in the ambulatory setting. Bupivacaine and levobupivacaine are used in epidural anesthesia when a longer block is desired. The use of 0.5% solutions is typical, but 0.75% solutions or adding epinephrine may provide increased motor block. Levobupivacaine is approximately equipotent to bupivacaine, with the exception of less systemic toxicity in animal models.

Ropivacaine was released for clinical use in 1996. When used as a spinal agent, ropivacaine has a clinical profile similar to bupivacaine at equipotent doses (ropivacaine is 60% as potent as bupivacaine) with little risk of TNS.30 Ropivacaine is used in concentrations of 0.5% to 1% for epidural anesthesia, providing blockade of somewhat shorter duration than bupivacaine. There is some evidence that ropivacaine may have less motor block when compared with bupivacaine when used for labor analgesia31; however, this effect is not reliably demonstrated when comparing equipotent doses.

Although tetracaine is used less commonly at this time, it remains the agent that provides the most reliable isobaric spinal anesthetic when the crystalline powder is reconstituted with the patient's cerebrospinal fluid. It should be noted that the use of vasoconstrictors might increase the risk of TNS with tetracaine.32

Investigators have modified use of long-acting agents as spinal anesthetics for short duration procedures by reducing the administered dose and thus reducing the duration of spinal blockade. When these agents are used for low-dose spinal anesthesia, this typically requires adjuvants (usually fentanyl) to have reliable block success.

Intermediate-Duration Agents

Lidocaine has enjoyed widespread popularity and perceived safety as a neuraxial agent, but it has undergone increasing scrutiny given the high incidence of TNS seen when used for ambulatory spinal anesthesia (see Neurologic Complications, later). In an attempt to avoid this problem, lower doses of lidocaine have been investigated, usually requiring adjuvant agents to provide suitable reliability of spinal blockade. Lidocaine is used in concentrations of 1.5% and 2% for epidural anesthesia to provide reliable blockade for procedures lasting less than 120 minutes. However, continuous catheter techniques commonly use lidocaine, with reinjection typically required every 60 to 90 minutes.

Mepivacaine has a clinical profile similar to that of lidocaine when used as a neuraxial agent, although it has higher potency (1.3:1 compared with spinal lidocaine).33 It also may have similar concerns regarding the high incidence of TNS when used for outpatient spinal anesthesia, depending on concentration of agent.34 It is used in concentrations of 1% to 2% for epidural anesthesia, again with a clinical profile similar to lidocaine.

Prilocaine is an amide local anesthetic with pharmacologic properties similar to those of lidocaine when used as a spinal anesthetic.35 However, it shows a much lower incidence of TNS compared with spinal lidocaine and thus may be a favorable agent for ambulatory spinal anesthesia. Because of the risk of methemoglobinemia at required doses, prilocaine is generally considered unsuitable for epidural anesthesia. Currently, prilocaine is not available in the United States, but it is used commonly in Europe.

Short-Duration Agents

As the first synthesized local anesthetic, procaine has been used for spinal anesthesia since the early 1900s. Procaine provides brief spinal anesthesia, but has limited clinical usefulness given the high incidence of block failure (see Table 47-4) and side effects. For reasons that are poorly understood, spinal procaine carries a higher risk of nausea than do other local anesthetics (odds ratio [OR] 3:1).36 Although it has a lower incidence of TNS than spinal lidocaine (Table 47-6), spinal procaine is clearly not enticing as an agent for outpatient spinal anesthesia.

2-Chloroprocaine has received increased attention as a spinal anesthetic because of the challenges of ambulatory spinal anesthesia and concerns of relative neurotoxicities of other local anesthetics (eg, lidocaine). One year after its introduction into clinical use (1952), Foldes and McNall described successful use of a preservative-free preparation of 2-chloroprocaine for spinal anesthesia in 214 patients. Subsequently, 2-chloroprocaine enjoyed increasing popularity as an agent for epidural anesthesia, particularly in the obstetric population. Unfortunately, reports of accidental intrathecal injection of large volumes of Nesacaine-CE (chloroprocaine hydrochloride; AstraZeneca, Mississauga, Ontario, Canada) intended for the epidural space, resulting in several cases of neurotoxicity with lower-extremity paralysis and sacral nerve dysfunction, came to attention in the 1980s. Studies in dogs suggest that the combination of the antioxidant sodium bisulfite in the presence of low pH was responsible for the neurotoxicity,37,38 and the formulation of 2-chloroprocaine was changed. Interestingly, a more recent laboratory study observed direct neurotoxicity from high doses of preservative-free 2-chloroprocaine in a rat model that was equivalent to 2% lidocaine and suggested that bisulfite was neuroprotective when added to 2-chloroprocaine.39 Currently, 2 of the commercially available formulations of 2-chloroprocaine (Nesacaine-MPF, Astra Pharmaceuticals, Wilmington, DE, and generic chloroprocaine, Bedford Pharmaceuticals, Bedford, OH) are preservative-free and antioxidant-free. Given the availability of these new preparations and growing concerns about the TNS associated with lidocaine, 2-chloroprocaine has been reinvestigated for off-label use as a short-acting spinal anesthetic. Work by Kopacz et al shows 2-chloroprocaine to be a reliable spinal anesthetic, with consistent time to block resolution and achievement of discharge criteria without identifiable occurrence of TNS.40-45 It should be noted that 2% 2-chloroprocaine may behave in a hyperbaric fashion,43 and that the addition of epinephrine to spinal 2-chloroprocaine is not recommended because of reports of flu-like symptoms in volunteers receiving this combination.40 As with fentanyl and plain bupivacaine, 2-chloroprocaine as a spinal anesthetic is currently an off-label use. However, the very low incidence of TNS and the rapid, dependable spinal block resolution is a very attractive profile for spinal anesthesia in ambulatory surgery. Given that the complete risk-to-benefit ratio of spinal 2-chloroprocaine is not known, clinicians should be mindful of restricting doses to not more than 40 mg, which is the most widely studied dose and ensure that only preservative-free, antioxidant-free preparations are utilized.

The use of 2% and 3% 2-chloroprocaine is an appropriate and US Food and Drug Administration (FDA)–approved choice for epidural anesthesia of short duration. After the release of Nesacaine-MPF, which contains ethylenediaminetetraacetic acid (EDTA), back pain following block resolution was reported with the use of high volumes of injected agent (>40 mL).47 One proposed mechanism for this observation is tetanic spasm of the paraspinous muscles resulting from chelation of calcium by the EDTA. However, this was never proved and reports of back pain persist, despite the use of EDTA-free solutions. Given the potential for any solution intended for the epidural space to be accidentally injected into the subarachnoid space, it may be prudent to use only 2-chloroprocaine preparations that are preservative- and antioxidant-free for neuraxial blockade.

Analgesic Adjuvants

Initially, interest in adjuvant medications for neuraxial anesthesia centered around increasing the duration and intensity of blockade (eg, epinephrine, phenylephrine). As the percentage of operations performed in the ambulatory setting increases, interest is shifting to identifying adjuncts that will augment block depth and reliability without prolonging recovery, especially motor block recovery (eg, fentanyl) (Table 47-7).

Opioids

The potent analgesic effects of neuraxial opioids have been exploited to improve perioperative analgesia and reduce the supraspinally mediated side effects of sedation and respiratory depression seen with systemic opioids. Neuraxial opioids that diffuse into the spinal cord exert spinal analgesia by modulating C-fibers to decrease afferent nociceptive input,48 inhibiting Ca2+ influx presynaptically, and increasing K+ conductance and hyperpolarizing ascending neurons postsynaptically.49

Owing to its relative hydrophilic nature, neuraxial morphine provides highly selective, prolonged spinal analgesia, but is not typically used to augment intraoperative anesthesia because of slow onset. The lipophilic opioids, such as fentanyl (note that intrathecal use is not FDA-approved), are more suited for intraoperative use in the intrathecal space because of rapid onset, modest duration, and lower risk of delayed respiratory depression (though greater risk of early respiratory depression). The rapid onset of fentanyl is due to multiple factors, including the relatively large dose typically used, noting that 25 μg of fentanyl is roughly equivalent to 2.5 mg of morphine. The addition of 10 to 25 μg fentanyl to low-dose lidocaine and bupivacaine spinal anesthetics dramatically improves anesthetic success without delaying achievement of discharge criteria for ambulatory patients.50,51 However, when used with the ultra–short-acting spinal anesthetic 2-chloroprocaine, fentanyl can slightly delay discharge (95 vs 104 minutes) and increase pruritus.42

The administration of epidural fentanyl can reduce volatile requirements more than intravenous fentanyl (>2-fold at 2 μg/kg).52 The method of delivery of epidural fentanyl may be important for optimal effect. Ginosar et al53 showed that when epidural fentanyl is given as a bolus, it imparts segmental analgesia consistent with spinal level of action, but if given as an epidural infusion, the analgesia is mediated through systemic uptake and supraspinal effect, as is seen with sufentanil and alfentanil.54

α2-Agonists

Interest in α2-agonists, such as clonidine, has been rising in the field of regional anesthesia given their ability to enhance neuraxial analgesia without the respiratory depression and pruritus common to opioids. As with analgesia, the sedation, hypotension, and bradycardia seen with neuraxial clonidine are dose dependent.55 Additionally, less urinary retention is seen with intrathecal clonidine than with intrathecal morphine.56 Clonidine exerts its analgesic effects by binding to α2-adrenoreceptors (on primary afferents, substantia gelatinosa, and several brainstem nuclei attributed to analgesic mechanisms), attenuating A-δ and C-fiber nociception, producing conduction blockade via increased potassium conductance,57,58 as well as by increasing acetylcholine and norepinephrine in the CSF, inhibiting the release of substance P.58 Although clonidine rapidly redistributes systemically to the periphery after epidural or spinal administration, the analgesic effect is spinally mediated, as evidenced by the lack of correlation between time of analgesia and peripheral blood levels. Through extensive testing for neurotoxicity and safety in several animal models, neuraxial clonidine shows no histopathologic or behavioral evidence of injury or toxicity and is FDA approved for epidural infusion for inadequately controlled cancer pain.

Although previously investigated as a sole anesthetic,59 the majority of the clinical use of intrathecal clonidine is in combination with local anesthetics to produce dose-dependent prolongation of both sensory and motor block.60-63 Showing a promising role in ambulatory anesthesia, De Kock et al64 demonstrated that the addition of as little as 15 μg of clonidine to 8 mg of ropivacaine for spinal anesthesia in outpatients undergoing knee arthroscopy produced a considerable increase in anesthetic success (from 70% to 90%) without significant effect on recovery time. However, increasing the dose to 45 μg increased time to resolution of motor and sensory block and time to void from 170 to 215 minutes. Adding clonidine to local anesthetics intensifies and prolongs epidural blockade and can reduce local anesthetic dose requirement.65,66 The typical dose of clonidine for addition to local anesthetics for epidural bolus administration is 150 μg, or 2 μg/kg.67-69 Klimscha et al68 demonstrated that the addition of 150 μg of clonidine to 10 mL of 0.5% bupivacaine for epidural anesthesia increased the mean duration of anesthesia from 1.8 to 5.3 hours, reduced pain scores, and increased time to first postoperative analgesic request. These benefits of clinical doses of neuraxial clonidine typically persist for approximately 3 hours and can be achieved without increasing hemodynamic instability more than local anesthetic alone or significantly altering responsiveness to resuscitation drugs.69-72

Vasoconstrictors

Vasoconstricting agents, namely epinephrine and phenylephrine, are commonly added to local anesthetic solutions and have a long history of clinical use to prolong the anesthetic effect, provide more reliable block, and intensify anesthesia and analgesia.73-76 Drugs with α-adrenergic activity appear to accentuate local anesthetic block by both pharmacokinetic and pharmacodynamic mechanisms. Pharmacokinetically, vasoconstriction of the arterioles in the dura mater,77 and thus decreased blood flow, can reduce uptake of local anesthetics into the circulation, thus maintaining concentrations at the site of injection and reducing peak plasma concentrations. Additionally, intrinsic analgesic effects of epinephrine are exerted via stimulation of presynaptic α2 adrenoreceptors found at the terminals of primary afferents. These receptors are also found centrally on neurons in the superficial laminae of the spinal cord and several brainstem nuclei that participate in analgesic mechanisms. α2-Agonists can also produce motor block via actions on primary motor afferents.78

Epinephrine

For spinal anesthesia, epinephrine is commonly used in a dose of 0.2 mg (although doses of 0.1-0.6 mg have been described), which, when added to a bupivacaine spinal anesthetic, increases time of regression to L2 by 25%.79,80 The addition of epinephrine to spinal anesthetics prolongs motor block and delays the return of bladder function, which is problematic for ambulatory surgery patients trying to meet discharge criteria. Chiu et al,81 using volunteers, showed that adding 0.2 mg of epinephrine to 50 mg of hyperbaric lidocaine prolonged surgical anesthesia (as demonstrated by tolerance of transcutaneous electrical stimulation) by 30 minutes, whereas time to void and discharge time were increased by 80 minutes.

As for safety, intrathecal epinephrine by itself, in clinically relevant doses, shows no neurotoxicity in humans. Spinal cord blood flow is well maintained in the dog and cat model in doses up to 0.5 mg.77 However, animal studies suggest that epinephrine may worsen the degree of injury associated with local anesthetic neurotoxicity.82 Smith et al40 reported consistent flu-like symptoms (myalgias, malaise, arthralgias, back stiffness, loss of appetite) when epinephrine was added to spinal 2-chloroprocaine in a volunteer study. Phenylephrine may increase the risk of TNS, as suggested by Sakura et al32 in a study of tetracaine spinal anesthesia.

With epidural anesthesia, the typical use of epinephrine is in concentrations of 1:200 000, or 5 μg/mL. The clinical effect of epinephrine on duration of anesthesia depends on the local anesthetic used. Epinephrine is more effective at prolonging the anesthetic duration of shorter-acting agents, such as lidocaine and 2-chloroprocaine. Adding 1:200 000 epinephrine to 2% lidocaine will nearly double the time to resolution of blockade.83 Agents with longer duration of action show much less prolongation of anesthesia with the addition of epinephrine. Adding epinephrine to ropivacaine will intensify the block, but will not prolong the duration of epidural anesthesia or affect plasma concentrations.74 This is likely a result of the inherent vasoconstricting effects of ropivacaine, as well as the clearance of ropivacaine not being dependent on blood flow. Other agents do show reduction of plasma concentrations when epinephrine is added.73,75 Epinephrine 1:200 000 will decrease plasma lidocaine and chloroprocaine concentrations by 20% to 30%, but will decrease plasma bupivacaine concentrations only by 10% to 20%. The effect of epinephrine on plasma concentrations of local anesthetics has long been thought to be caused by constriction of the epidural venous plexus, reducing blood flow and slower uptake of local anesthetics. More recent evidence implies that reduced dural blood flow and increased hepatic clearance may be more important in this phenomenon.84 Furthermore, work by Bernards et al85 suggests that systemic effects on vasculature of epidurally administered epinephrine may alter volume of distribution and thus contribute to the altered plasma concentrations of local anesthetics. Its potential to prolong discharge times and delay bladder function limits the usefulness of adding epinephrine to epidural agents for ambulatory surgery. The premixed solutions of local anesthetics with epinephrine that are commercially available are prepared more acidic to prevent spontaneous epinephrine oxidation. Because this lower pH slows the onset of block and inhibits the vasoconstricting actions of epinephrine, adding "fresh" epinephrine to local anesthetic solutions at the time of use is preferred. When phenylephrine is added to epidural solutions, the systemic absorption results in increased vascular resistance (as opposed to the decrease in systemic vascular resistance [SVR] seen with epidural epinephrine at 1:200 000 concentration), without the benefit of increased contractility or chronotropy seen with epinephrine. Consequently, phenylephrine is typically only used in the subarachnoid space.

Neostigmine

The acetylcholinesterase inhibitor neostigmine has been investigated as a neuraxial analgesic adjunct because of its ability to provide analgesia without hemodynamic depression. Unfortunately, its tendency to induce nausea and delay recovery from neuraxial blockade limits clinical use. Intrathecal neostigmine inhibits the breakdown of acetylcholine in the meninges and spinal cord via reversible inhibition of acetylcholinesterase. Animal models suggest that acetylcholine plays a role in spinal analgesia through stimulation of cholinergic receptors in the substantia gelatinosa and superficial laminae of the spinal cord dorsal horn86-88 and perhaps through stimulating nitric oxide production in the spinal cord.89 Whereas intrathecal injection of cholinergic agonists stimulates all receptors of a particular class, neostigmine increases endogenous acetylcholine in a manner dependent on the tonic release of this neurotransmitter within each particular region of the spinal cord. Hood et al90 evaluated safety, analgesic efficacy, and side effects of intrathecal neostigmine in volunteers. All doses produced analgesia without sedation, pruritus, respiratory depression, hypotension, or bradycardia; however, there was dose-related motor weakness, decreases in deep tendon reflexes, urinary incontinence, and nausea and vomiting. Further studies in patients revealed similar incidence of nausea and vomiting that proved to be both prolonged and difficult to treat.91-93 Liu et al93 showed that when added to low-dose (7.5 mg) bupivacaine spinal anesthetics, 50 μg of neostigmine enhanced motor and sensory block, but delayed achievement of discharge criteria. Doses of 6.25 and 12.5 μg did not prolong anesthesia but still elicited nausea and delayed discharge. Intrathecal neostigmine does counteract hypotension resulting from bupivacaine spinal anesthesia in rats,94 but these effects are not reproducible in human subjects.95 Low and moderate doses of neostigmine are considered to have little or no cardiovascular effects.

Lauretti et al96 studied the analgesic effect of epidural neostigmine, in doses from 1 to 4 μg/kg added to epidural lidocaine, and showed a dose-independent analgesic effect, increasing time to first analgesic request from 3.5 to 8 hours, with less nausea and vomiting than seen with intrathecal administration. Other studies report similar results with doses in the range of 1 to 10 μg/kg, without reporting increased nausea,97-102 but reporting some suggestion of sedation.101

Alkalinization and Carbonation

Alkalinizing local anesthetic solutions to raise the pH closer to the pKa of the local anesthetic, thereby increasing the proportion of the nonionized form available to cross cell membranes, is thought to speed the onset of epidural anesthesia. Although this is well demonstrated with peripheral nerves in vitro,103 studies attempting to demonstrate this clinically in epidural anesthesia are conflicting.

Most studies show that alkalinization speeds onset of epidural blockade with lidocaine,104-109 bupivacaine,104,110,111 mepivacaine,104,112,113 and chloroprocaine114,115 by up to 10 minutes. Ropivacaine seems not to show faster onset with alkalinization,116 but as with the other drugs, there is evidence that alkalinization can intensify epidural anesthesia and improve spread to sacral dermatomes.105,108,109,117 One trend that is noted is that the effects of alkalinization are greatest on solutions containing epinephrine, whether freshly added or prepackaged. This is perhaps a result of pH-dependent vasoconstrictive actions of epinephrine. Alternatively, this may be attributed to the fact that commercially available epinephrine-containing solutions are prepared at lower pH (usually with bisulfite), ranging from 3.2 to 4.2, in efforts to preserve the epinephrine.

Typically recommended volumes of 8.4% sodium bicarbonate to be added to local anesthetic solutions are 1 mL per each 10 mL of lidocaine or mepivacaine, 0.1 mL per 10 mL of bupivacaine, and 0.3 mL per 10 mL of 2-chloroprocaine. Because of the tendency to precipitate, adding sodium bicarbonate to ropivacaine solutions is not recommended. It should be noted that the degree of alkalinization is limited by precipitation, and all preparations should be inspected for precipitation before administration. Alternatively, the carbonate salts of local anesthetics have a more rapid onset of epidural blockade than standard hydrochloride preparations.118 However, carbonated drugs are of limited availability and may be more prone to induce hypotension with epidural administration.119

Neurophysiology

Neural blockade by local anesthetics is discussed in Chapter 45, and thus this section highlights some of the clinical aspects specific to neuraxial blockade. After intrathecal injection of local anesthetics, drug is found in the spinal cord as well as in the spinal nerve rootlets within the CSF. After epidural injection, local anesthetic is found in spinal nerves in the epidural space, spinal nerve rootlets within the CSF, and within the spinal cord. Consequently, conduction blockade may take place at multiple sites along the neural pathway for both spinal and epidural anesthesia, and the exact site(s) of action are not precisely known.120 However, previous studies do provide some insight into primary sites of action. After spinal anesthesia, somatosensory evoked potentials from the tibial nerve (peripheral nervous system) are abolished, whereas direct spinal cord stimulation remains unchanged.121,122 These findings lend support to the theory that the spinal nerve rootlets are the primary site of action of spinal anesthesia and not the spinal cord. Animal studies suggest that the mechanism of action of epidural anesthesia is similar to that of spinal anesthesia. Measurement of evoked potentials in monkeys again indicates that the primary site of action of epidural anesthesia is the spinal nerve rootlets.123 Interestingly, epidural anesthesia does not produce complete conduction block, as somatosensory evoked potentials from the tibial nerve are only modestly changed after induction of surgical epidural anesthesia. Furthermore, the magnitude of change in somatosensory evoked potentials does not correlate with intensity of epidural block.124 This finding is similar to peripheral nerve block, and investigators suggest that anesthesia occurs from loss of information coding inherent in the oscillations of the action potential instead of from complete loss of conduction. Finally, the ability of spinal but not epidural anesthesia to completely suppress somatosensory evoked potentials from the tibial nerve offers objective support for the clinical impression that spinal anesthesia produces a more intense and complete block than epidural anesthesia.

In addition to direct conduction blockade within the CNS, spinal and epidural anesthesia produce sedation that is unrelated to systemic concentrations of local anesthetic, but which correlates with block height.125 Animal studies using electroencephalogram (EEG) monitoring and direct brain stimulation during spinal anesthesia suggest that this is the result of a decrease in reticular activating system activity from decreased tonic afferent input from the anesthetized region.126 One may note decreased requirements for supplemental sedation during high spinal blockade from this direct sedative effect. Previous clinical studies examining use of propofol and midazolam suggest that spinal anesthesia per se decreases sedative requirements by approximately 30% to 50% with clinically relevant spinal block heights.127,128 This direct sedative effect of epidural anesthesia also becomes clinically relevant when a combined epidural-general anesthetic technique is performed. The use of epidural anesthesia reduces volatile anesthetic requirements by 20% to 30% during surgery when general anesthesia is titrated to either hemodynamics or bispectral index monitoring.129,130

Respiratory Physiology

With spinal or epidural blockade to midthoracic levels, pulmonary function, gas exchange, and control of breathing are generally preserved in patients without preexisting respiratory disease.131 Many patients will report a subjective sensation of dyspnea as a consequence of reduced sensation of expansion of the chest wall with inspiration. However, resting tidal volumes, respiratory rate, minute ventilation, and lung volumes are maintained in healthy patients.132,133 Preservation of gross pulmonary function—even with relatively high thoracic level blocks—is explained by the fact that the diaphragm is the primary muscle of ventilation and is innervated by the cervical plexus (C3-5). In contrast, accessory respiratory muscles (abdominal, intercostal) do play a role in active expiratory function, and small block height–dependent decreases in peak expiratory flow can be observed (11% reduction at T8 vs 17% reduction at T4).134 Active expiratory function plays a role in ability to cough; thus spinal anesthesia may impair ability to clear secretions.131 Overall, healthy patients easily tolerate these mild changes, but patients with severe pulmonary disease may not. However, previous studies indicate that patients with chronic lung disease do not suffer from significant reductions in vital capacity or forced expiratory volume at 1 second (FEV1) during spinal anesthesia.131

Control of breathing during spinal anesthesia is not altered significantly, although earlier studies demonstrated a small decrease in resting end-tidal Pco2.135 Although hyperventilation as a consequence of anxiety may cause lowering of the Pco2, the hypocapnia is speculated to result from a lack of proprioceptive input from the abdomen and chest wall during spinal anesthesia, resulting in an increased drive to breathe.135 It has been reported that spinal anesthesia with bupivacaine in unpremedicated patients increases ventilatory responsiveness to CO2.136 The rare respiratory arrest after spinal anesthesia is thought to result from brainstem hypoperfusion secondary to decreased cardiac output rather than the direct effects of local anesthetics on the brainstem, as the concentration of local anesthetic in the ventricular fluid is not high enough to result in medullary depression.

Preoperative discussions usually encourage tolerance of the subjective dyspnea, but at times judicious sedation can be used to quell anxieties. It should be noted that sedative medications used to facilitate neuraxial blockade are more likely to impact the patient's respiratory status than the blockade itself.140 This is likely a result of both reduction in respiratory drive and initiation of paradoxical respiration from upper airway obstruction.

Cardiovascular Physiology

The perturbations of cardiovascular function during neuraxial anesthesia are perhaps the most critical factors to consider in evaluating the risk to patients and in preventing adverse outcomes. In the nonobstetrical population, the incidences of hypotension and bradycardia after spinal anesthesia are approximately 33% and 13%, respectively.28,119 Large epidemiology studies from France, Scandinavia, and the United States indicate that the risk of cardiac arrest after spinal anesthesia is approximately 0.1-1:1000 and 1:10 000 after epidural anesthesia.137 Although cardiac arrest after spinal anesthesia appears to be distressingly common, a recent study suggests that survival is better for cardiac arrest during neuraxial block than during general anesthesia (65% vs 31%).138 This may be a result of enhanced vigilance, as several risk factors for bradycardia and hypotension have been identified for spinal anesthesia (Table 47-8). Risk factors for epidural anesthesia are probably similar but have not been fully identified because of the decreased frequency of occurrences. An appreciation for the mechanisms involved in these physiologic derangements allows early recognition and prompt treatment of potentially detrimental situations.

Cardiovascular changes seen with spinal anesthesia are caused by blockade of the sympathetic efferent fibers and thus generally are related to block height.139 Both arterial and venous relaxations contribute to hypotension, resulting from decreases in SVR and cardiac output. SVR decreases to a greater degree in patients aged 69 to 80 years (26% from baseline) than in young, healthy subjects (13%-18% from baseline).140 Venodilation causes increased pooling of blood in the capacitance vessels, thus reducing central blood volume.

Although heart rate is typically maintained, bradycardia can occur with spinal anesthesia in 10% to 15% of cases, and unexpected circulatory collapse remains a dreaded complication with a potentially grave outcome. In addition to the risk factors of age younger than 50 years, American Society of Anesthesiologists (ASA) physical status P1, and concomitant use of β-blockers, the incidence of bradycardia increases with increased block height with 75% of occurrences associated with sensory block above T536 (Table 47-8). Cardioaccelerator fibers originate from T1 to T5, and thus sympathetic blockade above T5 is thought to allow parasympathetic predominance over heart rate, mediated via the vagus nerve. However, some studies of heart rate variability demonstrate that even with high thoracic blockade, the activity of sympathetic and parasympathetic systems can remain in balance.141,142 Reports of bradycardia/asystole and circulatory collapse in patients with blocks too low to be attributed solely to sympathectomy are further evidence that other factors may play important roles. Bradycardia may be induced by an increase in baroreceptor reflex activity.143 With redistribution of blood to capacitance vessels and decreased venous return to the heart, resulting in decreased filling pressures, intracardiac stretch receptors within the right atrium and left ventricle have been suggested to participate in a bradycardic response (Bezold-Jarisch reflex; Fig. 47-14).144,145 Maintaining preload and aggressive treatment of bradycardia may help improve the safety of spinal anesthesia.146

Efforts have been made to investigate the usefulness of heart rate variability analysis for predicting hypotension after spinal anesthesia for cesarean section to assess the underlying balance of patients' autonomic systems.147,148 Although small and limited, these studies demonstrate that measurements of heart rate variability are predictive of hypotension in this patient population. Heart rate variability is extracted from the patient's electrocardiogram (ECG) and analyzed based on the preponderance of low versus high frequencies to determine subsets of patients who are at risk and who would benefit from additional prophylactic interventions (see next paragraph). Computer-aided analysis is being developed, as currently this is not a common ECG analysis. More research is needed, but this appears to be a promising clinical tool to direct selective therapy aimed at preventing hypotension with spinal anesthesia.

The prevention and prompt treatment of hypotension and bradycardia during spinal anesthesia is essential in protecting the patient from untoward outcomes. The practice of administering a crystalloid bolus before spinal anesthesia has been a standard method for preventing hypotension induced by neuraxial blockade. Behind this is an effort to maintain central blood volume, and, therefore, venous return to the heart, so as to preserve cardiac output. However, studies reveal a more complicated interaction between fluid loading, hemodynamic effects, and efficacy for preventing hypotension in the nonobstetric population. For example, prophylactic bolusing of 500 to 1500 mL of crystalloid may be ineffective for prevention of hypotension in normovolemic patients if performed before induction of spinal anesthesia,149-151 but may be effective if performed later during the actual performance of spinal anesthesia.152 This may be a result of the rapid redistribution of crystalloids out of the intravascular compartment, thus providing only a fleeting contribution to venous return.153 Furthermore, a crystalloid bolus does not adequately address other factors contributing to hypotension, namely heart rate and SVR, and may actually decrease SVR.154 In contrast, the administration of colloid solutions is more effective in maintaining intravascular volume caused by favorable pharmacokinetics154,155 and may actually increase SVR.154 Prophylactic administration of 500 to 1000 mL of colloid solution before induction of spinal blockade more effectively prevents hypotension but can grossly affect fluid balance in the patient. In regards to treatment of hypotension in the setting of established hypovolemia, crystalloid remains a suitable choice because of altered volume kinetics in such a setting.156

The technique of spinal anesthesia may also be altered to attenuate resultant vasodilatation and hypotension. Sympathetic block is somewhat dose dependent; thus selection of an appropriate dose is helpful. The use of unilateral spinal anesthesia for unilateral lower-extremity procedures allows limited spread of sympathetic block and has been investigated. Typically, a small dose of hyperbaric local anesthetic is injected and the patient is kept in the lateral position (operative side down) for 15 minutes. Casati et al157 showed a decreased rate of hypotension in patients with intentionally asymmetric spinal blocks as compared with conventional bilateral blockade (5% vs 22%). Incidence of hypotension is reduced with epidural compared with spinal anesthesia, and placement of a catheter with gradual titration of epidural local anesthetic dose allows time for treatment of the gradual sympathectomy.

Effective treatment of hypotension should be tailored to the clinical situation at hand, with consideration of alterations in SVR and cardiac output. Vasopressors with α-adrenergic agonist activity, such as phenylephrine and metaraminol, are quite effective for increasing SVR. However, this may be at the expense of a further decrease in cardiac output as a response to the increase in afterload.158 In light of this, it is thought that the use of mixed α- and β-adrenergic agonist drugs, like ephedrine, may be more appropriate for the treatment of hypotension induced by neuraxial blockade as a result of the ability to augment heart rate and cardiac output as well as SVR. Ephedrine is most often administered as an intermittent intravenous bolus of 5 to 10 mg, but may also be given via continuous infusion or intramuscularly as a depot of 25 to 50 mg. Atropine in doses of 0.4 to 1 mg intravenously can be used to treat moderate bradycardia if ephedrine is ineffective. In the case of precipitous bradycardia, or in situations unresponsive to the previously mentioned interventions, one should not hesitate to consider administration of epinephrine. Although clinical data are lacking, closed claims data analysis suggests that the lack of early administration of epinephrine is a management pattern in spinal anesthesia that leads to cardiac arrest with poor outcome.159

Although similar hemodynamic changes are seen with epidural anesthesia, these tend to be better tolerated than changes seen with spinal anesthesia, perhaps because of a more gradual and titratable onset. Nonetheless, sudden bradycardiac cardiac arrest has been described with epidural anesthesia.160 Risk of sudden cardiovascular collapse from epidural anesthesia may be influenced by the addition of epinephrine to the epidural solution. With the typical dose range used for epidural anesthesia, systemic levels of epinephrine remain low, producing a β2-adrenergic effect of vasodilatation, increased heart rate, and myocardial contractility. Ward et al161 evaluated the cardiovascular effects of epidural blockade to T5 using lidocaine with and without epinephrine. Mean arterial pressure decreased 20% in the epinephrine group, compared with 10% in the plain lidocaine group. However, the group with epinephrine also showed a 20% to 30% increase in cardiac output. Bonica et al 162 suggested that this systemic β-adrenergic effect of epinephrine might prevent the potential cardiovascular collapse from epidural blockade, though this has not been demonstrated in studies of sufficient power.

Gastrointestinal, Hepatic, and Genitourinary Physiology

The sympathectomy of spinal anesthesia results in a relaxation of sphincters, constriction of the bowel, and an increase in secretions caused by "parasympathetic dominance." This imbalance of the autonomic nervous system is also thought by some to explain the occurrence of nausea seen with spinal anesthesia. Hepatic blood flow is related to mean arterial pressure and is thus maintained if the patient is hemodynamically stable.163 Likewise, renal blood flow and renal function are preserved during spinal anesthesia when perfusion pressure is adequate.164 Urinary retention after spinal anesthesia is the most noteworthy and clinically significant concern in regard to the genitourinary system. Postoperative urinary retention occurs in approximately 16% of patients in the recovery unit.165

After the induction of spinal anesthesia, the urge to void (normal detrusor function) is abolished within 60 seconds.166,167 Recovery of the ability to void normally does not return until sensory anesthesia has regressed to the S3 sacral segment.166 Prolonged inhibition of normal detrusor function with the use of long-acting local anesthetics such as bupivacaine may allow bladder overdistension and urinary retention. In a study of healthy male patients undergoing nonurologic surgery after spinal anesthesia comparing 100 mg lidocaine with 10 mg bupivacaine, the time to return of normal detrusor function was significantly longer in the bupivacaine group (233 ± 31 vs 462 ± 61 minutes).166 The cystometric capacity (the bladder volume at which patients feel an urge to void before spinal anesthesia) in this study was between 500 and 600 mL, and the patients in the bupivacaine group generated an average of 875 ± 385 mL of urine, far exceeding the cystometric capacity and suggesting that the use of long-acting local anesthetics may lead to bladder overdistension and urinary retention. Furthermore, the use of epinephrine as an adjuvant for spinal anesthesia has been identified as a factor in delayed voiding ability.81 Other factors such as age (>50 years), volume of intraoperative fluid administration, and type of surgical procedure also influence rate of urinary retention.165

Concern for postoperative urinary retention is especially important in the ambulatory setting, where the traditional requirement to void after spinal anesthesia often leads to prolonged delays in discharge.168 However, optimal use of short-acting local anesthetics for ambulatory spinal and epidural anesthesia has not been associated with urinary retention.169 A study of ambulatory surgery patients discharged before voiding after short-acting spinal anesthesia demonstrated significantly shorter discharge times, with no reports of urinary retention.170 Thus the risk of urinary retention appears to be low after short-acting spinal anesthesia, and further prospective study is needed to confirm this practice (discharging patients before voiding after short-acting spinal anesthesia) in a large population.

Postmenigeal Puncture Headache

Appreciated since perhaps the first spinal anesthetic, postmeningeal puncture headache (PMPH) remains a complication of spinal anesthesia and is seen in as much as 50% of cases of unintentional meningeal puncture during attempted epidural anesthesia. Reported incidence of PMPH after spinal anesthesia varies greatly, depending on the types of needles used and patient population. Although rates as high as 40% are seen with 22-gauge beveled-tip needles,171 the use of smaller-diameter pencil-point needles has decreased the risk of PMPH after spinal anesthesia to approximately 1%.172 The traditional mechanism of PMPH is believed to involve the loss of CSF from the thecal sac and a "sagging" of the brain while standing or sitting upright. This is thought to result in traction on the cranial meninges, meningeal vessels, and, at times, traction on the cranial nerves, leading to cranial nerve palsy. Magnetic resonance imaging (MRI) studies demonstrating reduced CSF volume173 and meningeal enhancement postgadolinium during PMPH174 lend support to this theory of CSF loss and meningeal irritation. There are also data suggesting that vasodilatation of cerebral vessels may be important in etiology of PMPH.175 A persistent hole in the dura mater has traditionally been blamed for PMPH, but there is no evidence that the dural hole is more important than the arachnoid hole in this regard. Consequently, we use the more general term postmeningeal puncture headache.

The headache is typically positional, being most severe while standing or sitting and nearly eliminated when supine. Patients usually describe a band-like aching pain in the frontal and occipital regions, posterior neck pain, and at times nausea, tinnitus, photophobia, and diplopia. Differentiation from infectious meningitis or intracranial mass lesions is important and usually based on positional nature of symptoms, absence of lateralizing neurologic findings, lack of fever, and, if necessary, normal peripheral white blood cell count and CSF profile.

Although traditionally held to the contrary, patient sex and postoperative recumbency are not related to incidence of PMPH.176 Although the incidence of PMPH decreases with increasing patient age,176 factors that are in the control of the clinician to reduce the occurrence of PMPH include the use of smaller-gauge needles, pencil-point needles, and longitudinal orientation of the bevel if a cutting needle is used.176,177 The mechanisms in play for the latter 2 factors are not entirely clear. The classically described scenario is that of more dural fibers being cut by a transversely oriented cutting bevel and thus a more substantial hole in the dura. However, given that the dural fibers are not actually longitudinally arranged, rather, randomly, the number of these fibers cut will not depend on bevel orientation. It has been suggested that the longitudinal tension placed on the meninges tends to pull open a transversely oriented defect and thus allow for more CSF leakage. As for pencil-point needles, the supposition that they cause less trauma to the meninges is questionable. Reina et al173 suggest that Whitacre needles produce more trauma, as evidenced by electron microscopy images, and that the reduced loss of CSF may be a result of an "edematous plug" resulting from greater inflammatory reaction.

Conservative management of PMPH includes bed rest, oral analgesics, adequate hydration, and caffeine (because of its ability to constrict cerebral arterioles). Although these measures are often ineffective, patients should be reassured that the symptoms are likely to resolve within 1 week without further invasive treatments. If the patient is not responding adequately to conservative treatments or is having prolonged symptoms, then an epidural blood patch can be considered. This is usually performed by accessing the epidural space within 1 interspace of the suspected meningeal hole and injecting 10 to 20 mL of autologous blood in an aseptic fashion. This is intended to form a clot, or "patch," over the dural defect, preventing further leakage of CSF. Additionally, this volume will tamponade the dural sac and restore buoyant support to the brain, explaining the near-immediate relief that many patients report. The success of epidural blood patch is reported to be in the range of 70% to 95%, with those who fail to improve with an initial patch showing the same range of response to a repeat procedure. Large needle causing the meningeal puncture and early administration of the treatment (delay of <4 days from puncture) are factors suggested to be associated with failure of epidural blood patch.178 The effectiveness of prophylactic blood patch administration is controversial, with some studies showing no benefit and others reporting greater than 50% success in preventing PMPH.179,180 More recent evaluations demonstrate that prophylactic blood patch will not prevent PMPH, but might at best shorten the duration of symptoms in obstetric patients with inadvertent meningeal puncture with a large-diameter needle.181

In the search for alternatives to epidural blood patch, administration of saline182 and dextran183 has been performed, but without prolonged relief of symptoms. Other substances, such as fibrin glue, have been used in the epidural space with some success,184,185 but further investigation is needed to determine the potential benefits and safety of such procedures.

Infectious

Infectious complications after neuraxial anesthesia are rare, but can arise in the form of meningitis and epidural abscess.186,187 In a large and relatively complete retrospective review, Moen et al188 estimated the incidence of meningitis to be less than 1 in 50 000 after spinal anesthesia and approximately 1 in 90 000 after epidural procedures and the incidence of epidural abscess to be 1 in 37 000 after epidural blocks. The source of microorganisms can be from contaminated equipment or injected solutions, or from patient source, namely bacteremia. The concern for potential inoculation from the performer's oral or nasopharyngeal flora has led to a CDC recommendation and the American Society of Anesthesiology Practice Advisory regarding the use of masks covering nose and mouth when performing neuraxial techniques.187 It is speculated that lumbar puncture may disrupt the blood–brain barrier and allow transfer of blood-borne bacteria into the spinal space. Animal studies support this theory, but also show pretreatment with antibiotics to drastically reduce this risk.189

Epidural anesthesia for surgical procedures may have a similar risk profile, but the use of indwelling epidural catheters for continuous analgesic infusions provides an additional risk of serving as a wick for surface infections to migrate along the tract to become a deep infection or epidural abscess. Although some surveillance studies and studies examining rates of bacterial contamination of epidural catheter tips indicate that risk is low,188,190 a survey by Wang et al191 reported an incidence of 1 in 1930 catheters, with prolonged duration of catheterization identified as a risk factor. Patients with postoperative catheters should be under surveillance for signs of infection, having the catheter site inspected daily. If signs of skin infection are present, the catheter should be removed and the site monitored for improvement, with or without the initiation of antibiotics as deemed suitable. If the patient presents with severe back pain, and/or new neurologic deficits that are not explained by the analgesic infusion, then the diagnosis of epidural abscess must be considered, and MRI or computed tomography (CT) performed if appropriate. Du Pen et al192 showed successful nonsurgical management of infections associated with chronic tunneled epidural catheters. However, in the setting of neurologic compromise, prompt diagnosis and surgical evacuation is imperative to permit recovery if an epidural abscess is identified.

Hemorrhagic Complications

With any invasive procedure, the possibility of inducing bleeding, especially in a noncompressible region, enters the risk-to-benefit analysis. With neuraxial procedures, hematoma can develop in the epidural, subdural, or intrathecal space, with the potential for devastating neurologic outcome. The estimated incidence of spinal hematoma in the absence of anticoagulant and antiplatelet medications is less than 1 in 150 000 neuraxial blocks,193 although Moen and colleagues194 identified significantly higher rates in elderly females and patients with concurrent coagulation-modifying factors. The majority of cases of spinal hematoma are in patients receiving anticoagulants or with hemostatic abnormalities from hepatic dysfunction or thrombocytopenia.

The usual presenting signs/symptoms are motor or sensory deficits or bowel/bladder dysfunction that are not explained by the administered anesthetic/analgesic agents. In patients with such symptoms, especially in the setting of anticoagulants, definitive diagnosis should be sought without delay. A review by Vandermeulen et al195 indicated that early identification by MRI or CT and surgical decompression performed within 4 to 8 hours of presentation is associated with good neurologic recovery. This review also highlighted the importance of coagulation status at the time of epidural catheter removal, with nearly 50% of the epidural hematomas in patients with indwelling catheters manifesting clinical symptoms after removal of the catheter.

The American Society of Regional Anesthesia and Pain Medicine has released an updated consensus statement to address the concerning issues surrounding neuraxial procedures in patients receiving anticoagulant and antiplatelet medications.193 Some of the recommendations from this statement are presented here, but this in no way substitutes for a full review of the information provided by the Consensus Conference, which can be downloaded for free at www.asra.com. One should also consider that any patient who is receiving combinations of the following medications is likely at further increased risk and appropriate caution taken.

Warfarin therapy should be stopped 5 days before planned intervention and international normalized ratio (INR) measured before neuraxial procedures. In patients with indwelling epidural catheters who have had warfarin therapy initiated, a lower-extremity neurologic examination protocol should be followed, and the INR should be confirmed to be less than 1.5 before epidural catheter removal.

The use of aspirin and nonsteroidal anti-inflammatory drugs is not considered to increase the risk of hemorrhagic complications with neuraxial procedures. However, clopidogrel should be discontinued for at least 7 days before neuraxial techniques. Given shorter half-lives, the glycoprotein (GP) IIb/IIIa antagonists only need to be held for 24 to 48 hours to allow return of platelet function.

Patients receiving twice-daily 5000 units of subcutaneous heparin for thromboprophylaxis are not considered to be at increased risk for spinal hematoma, but considerations should be made with regard to timing of procedures in light of the peak onset of 2 hours after administration of heparin dose. The safety of subcutaneous heparin administration in a 3-times daily regimen or in larger doses has not been determined. Intraoperative anticoagulation with IV heparin is acceptable if given 1 hour after and discontinued 2 to 4 hours (and normal partial thromboplastin time [PTT] confirmed) before block placement or catheter removal.

Low-molecular-weight heparins (LMWHs) present added management concerns given the prolonged action (especially in patients with decreased renal clearance) and difficulty in measuring the induced alterations in coagulation status. Current recommendations state that patients receiving thromboprophylactic doses should have neuraxial procedures delayed 10 to 12 hours after last dose. If receiving "treatment" dosing (1 mg/kg twice a day), a delay of 24 hours is recommended. After neuraxial procedures, initiation of LMWH therapy should be delayed 6 to 8 hours for daily dosing or 24 hours for twice-daily dosing.

Neurologic Complications

Neurologic injury from neuraxial procedures is rare, but it remains a fervently held fear among many patients. The overall incidence of persistent neurologic injury associated with neuraxial blocks is reported to be approximately 0.08% to 0.16%; however, the vast majority of these reported cases fail to show evidence that the block was directly causative.196 Potential mechanisms of neurologic injury after neuraxial anesthesia include direct needle or catheter trauma, neurotoxicity of injected substances (intended agents or unintended chemicals), infectious complications, hemorrhagic complications, or spinal cord ischemia. The majority of perceived injuries are related to persistent paresthesia or motor weakness, but cauda equina syndrome is also rarely reported.

Although paresthesia is reported in as many as 13.6% of patients during needle placement for spinal anesthesia,197 this rarely results in persistent neurologic deficit.198 Similarly, most paresthesias associated with indwelling epidural catheters are likely to dissipate after removal of the catheter.199 Despite its rare occurrence, paresthesia during block placement is considered a risk factor for persistent paresthesia. Patients experiencing this should be reassured and queried postoperatively regarding their neurologic status.

Attention has been focused on the issues surrounding neurotoxicity of local anesthetics and common additives because of identification of several cases of cauda equina syndrome associated with continuous spinal anesthesia via microcatheters. Although the injuries in these cases have been attributed to supranormal doses of local anesthetic, pooling and maldistribution of the local anesthetic within the thecal sac caused by the nature of administration through the small diameter catheter,200 there are reports of similar injuries from single-shot lidocaine spinal anesthetics using relatively high doses (>75 mg) with epinephrine.201 Despite the long history of relative safety, animal data exist to suggest that all local anesthetics have some potential to cause neural injury; however, lidocaine and tetracaine appear to have a greater potential for neurotoxicity than bupivacaine at clinically relevant concentrations.202 Drasner203 has presented recommendations for lidocaine spinal anesthesia, including limiting the dose to 60 mg, avoiding epinephrine, and restricting concentrations to 2.5% or less, although little human evidence exists to support these proposals. The concerns of dose, concentration, and adjuvants might be applicable to all local anesthetics.

A much more commonly reported complication after neuraxial anesthesia is transient neurologic symptoms (TNS). These symptoms, initially termed transient radicular irritation in 1993,204 are described as back pain radiating to the buttocks and/or the lower extremities. Although TNS has been reported after spinal anesthesia with all local anesthetics, it is far more common with lidocaine, showing an incidence in prospective studies to be between 4% and 36%.29 Drastic variation in the observed rates of TNS (Table 47-6) lends support to the theory that multiple factors are involved in its development. Factors demonstrated to contribute to the incidence of TNS include the use of lidocaine (relative risk of 4.35 compared with other local anesthetics),205 lithotomy position, knee arthroscopy, and early ambulation (as in outpatient surgery) (Table 47-6).204

The etiology of TNS remains elusive, but proposed culprits include direct neurotoxicity of local anesthetics, stretching of neural structures via positioning and muscle relaxation, and muscular spasm. The evidence against neurotoxicity being the principal cause of TNS includes lack of motor deficit or electrophysiologic changes during acute symptoms,206 as well as its response to nonsteroidal anti-inflammatory drugs and other modalities targeting muscular discomfort.

Local Anesthetic Systemic Toxicity and Epidural Test Dose

It is important to verify correct placement of needles and catheters in the epidural space before initiating epidural anesthesia. Unintentional subdural or subarachnoid dosing can result in high blocks or "total spinals," whereas intravascular injection can result in local anesthetic systemic toxicity (LAST), which can progress to seizures or cardiovascular collapse.207 Subdural injection is rare and may be difficult to recognize, as CSF will not be aspirated from the needle or catheter. The clinical pattern is one of an unusually high block after administration of local anesthetic.208 Unintentional subarachnoid injection is a more common occurrence with incidences of "total spinals" ranging from 0.3% to 0.6%. It is important to realize that an epidural catheter may migrate into the subarachnoid space at any time during use. Unintentional intravascular injection is probably more common, with incidences of systemic toxic reactions (ranging from mild CNS symptoms to seizures) ranging from 0.01% to 2%.208 Again, it is important to realize that an epidural catheter may migrate into the intravascular space at any time during use.

Although efficacy for improving safety is uncertain, it is standard practice to administer a test dose before initiating full epidural anesthesia.209 This test dose is intended to detect unintentional placement of the needle or catheter into either the subarachnoid or intravascular space. A 3-mL volume of local anesthetic is the primary component of a test dose. Theoretically, subarachnoid injection of this local anesthetic will produce a much more rapid and profound sensory and motor block than epidural injection. However, changes can be subtle and will require time to develop. Previous reports indicate that 2 to 4 minutes are required before development of sensory or motor block after subarachnoid injection of 3 mL of 1.5% lidocaine.208,210 Subarachnoid injection of a bupivacaine test dose (3 mL 0.25%-0.5%) produces spinal blocks with highly variable onset time and spread and is probably not a reliable indicator.208 Use of a test dose to detect subarachnoid placement is not without risk, as previous publications have reported high spinal anesthesia requiring tracheal intubation after 3 mL of 1.5% lidocaine or 0.5% bupivacaine.208 The local anesthetic component of the test dose has been suggested to aid in detection of intravascular injection by producing classic symptoms of systemic toxicity, such as tinnitus, perioral tingling, metallic taste, and dizziness; however, the presenting symptoms of LAST are variable and possibly delayed.211 Detection ability is limited, as the standard 3-mL test dose (45 mg of lidocaine or 15 mg of bupivacaine) contains insufficient local anesthetic to reliably produce such symptoms. Previous studies suggest that larger doses, such as 100 mg or 1 mg/kg of lidocaine or more than 25 mg of bupivacaine, levobupivacaine, or ropivacaine, are required in the unsedated patient.208,212 Administration of even modest doses of sedation further reduces the ability of patients to report these symptoms (40% reduction in sensitivity).213

Epinephrine is commonly added to the local anesthetic dose to increase sensitivity for detection of intravascular injection. Addition of 15 μg is the standard dose of epinephrine; it will consistently produce increases in heart rate and systolic blood pressure and reduction in T-wave amplitude within 60 to 120 seconds in healthy patients (Table 46-9).208,214 Use of β-blockers, advanced age, and concurrent general or spinal anesthesia all decrease the hemodynamic response to this dose of epinephrine, and reduced criteria, or larger epinephrine doses, should be applied in these circumstances.208,215,216 Intravascular injection of epinephrine may be of concern in patients with hypertension, patients at risk for myocardial ischemia, and in obstetrical patients, and potential risks should be considered in these patients.

The best method for avoiding systemic toxicity from local anesthetics is through prevention.207,209 Toxic systemic concentrations can occur by unintentional intravenous or intra-arterial injection or by systemic absorption of excessive doses placed in the correct area. Unintentional intravascular injections can be minimized by frequent syringe aspiration for blood, use of a local anesthetic test dose (see earlier discussion), and either slow injection or fractionation of the rest of the dose of local anesthetic.209 Detailed knowledge of local anesthetic pharmacokinetics will also aid in reducing the administration of excessive doses of local anesthetics. Ideally, heart rate, blood pressure, and ECG should be monitored during administration of local anesthetics. Pretreatment with a benzodiazepine might also lower the probability of seizure by raising the seizure threshold. The administration of intravenous lipid emulsion in the treatment of cardiovascular toxicity has gained much attention after demonstrated benefit in experimental animals and case reports in human patients.217 Although acceptance of this treatment is rapidly taking hold in clinical practice, further evaluation is needed to identify optimal administration and potential adverse effects. A treatment algorithm for LAST can be accessed for free at www.asra.com through the Practice Advisory section.

Because of the density and reliability of spinal anesthetic blockade, attempts to extend the duration and graduate the dosing of agents led to techniques for continuous spinal anesthesia (CSA). A century ago, malleable spinal needles were left within the subarachnoid space to allow redosing during prolonged procedures, and later, Tuohy placed catheters in the intrathecal space. The development of microcatheters (29-32 gauge) in the 1980s led to a revisiting of this technique, with hopes of titrating and prolonging spinal blockade with reduced incidence of PMPH as compared with Tuohy's 15-gauge catheters. Unfortunately, beyond the problems of breakage and knotting, these catheters are associated with several reports of cauda equina syndrome, usually with 5% hyperbaric lidocaine used as the agent. The hypothesis is that the small-bore catheters reduced local anesthetic mixing in the CSF, which in turn led to a "sacral pooling" of the hyperbaric agents, limiting spread and dilution of the drug in the CSF, unmasking the neurotoxic potential of lidocaine at such a concentration. The FDA subsequently removed these microcatheters from the market, as did the Canadian authorities. However, a recent evaluation of bupivacaine and sufentanil via 28-gauge intrathecal catheters by Arkoosh et al218 encourages renewed interest in microcatheters, demonstrating superior analgesia without evidence of neurologic complications, but with increased technical difficulty with placement and removal. Although some practitioners still use continuous spinal anesthesia using larger-gauge catheters, the risks and benefits must be weighed carefully for the given situation. If CSA is to be undertaken, precautions should be taken to not advance the catheter beyond 3 cm within the intrathecal space, avoid dextrose-containing preparations, give adequate time for spread and distribution of the agent befre additional dosing, and be prepared to abandon the technique if sacral pooling is suspected.

Combined spinal–epidural anesthesia (CSEA) is an increasingly popular technique owing to its advantages of rapid onset, dense neuraxial block, ability to titrate spread and duration of block, and lower total drug dosage when compared with epidural anesthesia. Potential disadvantages include increased time to perform the dual technique, intrathecal migration of epidural drug and/or catheter, effects of increased epidural pressure from injection of solutions on spinal block, and decreased ability and reliability of epidural test dosing. Although a seemingly simple combination of 2 routine techniques for neuraxial anesthesia, the interactions between the 2 can be subtle, can impact clinical management, and have not been fully determined.

Patient Selection and Clinical Applications

Obstetrics

CSEA has been most widely accepted in the obstetric population. The concept of the "walking epidural" has become popular among patients, where intrathecal opioid allows rapid onset of analgesia without motor blockade. Lipid-soluble opioids, such as fentanyl (up to 25 μg) and sufentanil (up to 10 μg), are commonly used to provide dose-dependent analgesia for 60 to 90 minutes. Opioids are also commonly combined with small doses of local anesthetics, such as bupivacaine, to either prolong this initial spinal analgesia or to reduce side effects by decreasing the required dose of opioids. Previous dose–response studies indicate that 2.5 mg of bupivacaine combined with either 15 μg of fentanyl or 2.5 μg of sufentanil provides satisfactory analgesia while reducing incidences of nausea and pruritus, when compared with larger doses of opioid.219,220 Preservation of motor function is an important goal after CSEA for labor analgesia, and sophisticated testing of positional sense and motor function after CSEA indicates little effect from the initial spinal dose (2.5 mg of bupivacaine + 5 μg of fentanyl).221 However, subsequent injection of a standard test dose (45 mg of lidocaine + 15 μg of epinephrine) to confirm epidural catheter placement or the initiation of continuous epidural analgesia appears to degrade motor function and positional sense.221,222 Addition of other less commonly used adjuncts to standard doses of opioids can also increase duration of analgesia. For example, addition of 200 μg of epinephrine or 50 μg of clonidine to opioid is equivalent to adding 2.5 mg of bupivacaine and can provide an additional 30 minutes of analgesia.223 Despite the apparent preservation of motor function, large clinical trials have not observed a decrease in incidence of cesarean section when CSEA is compared with conventional epidural analgesia, although the incidence is similar to systemic analgesia.224-226

As an anesthetic for cesarean section, CSEA offers a rapid, titratable block, good muscle relaxation if an appropriate local anesthetic concentration is used, and the ability to use reduced doses of local anesthetic. Several clinical trials have compared CSEA with epidural anesthesia with lidocaine or bupivacaine combined with fentanyl. These studies report more rapid onset, better motor block, decreased anxiety levels, decreased shivering, and greater patient satisfaction with CSEA.227,228 Incidences of side effects were similar in that the severity of hypotension did not differ, nor did the incidence of postmenigeal puncture headaches, backaches, nausea, or vomiting.227,228 When compared with conventional single-shot spinal anesthesia, use of CSEA offers the potential advantage of using a smaller initial dose of spinal anesthetic with the epidural catheter available, should the block be insufficient. Such an approach theoretically offers the advantage of greater individual titration with reduced side effects and faster recovery from conduction block. However, results from studies are inconsistent, with some trials observing faster motor recovery with this approach229 and others not discerning a reduced dosing need for spinal anesthesia with epidural supplement.230,231 A potential disadvantage of the CSEA technique versus the single-shot technique is an increased incidence of transient paresthesia during placement of the spinal needle. A previous study randomizing patients undergoing cesarean section to CSEA versus conventional spinal anesthesia observed incidences of paresthesia of 37% versus 9%. There were no long-term complications, and the authors speculated that the CSEA technique might lead to deeper tissue penetration as a mechanism for increased incidence of transient paresthesia.232

Ambulatory Anesthesia

The dose of local anesthetic determines both anesthetic success and duration of recovery. Availability of the epidural catheter for a rescue anesthetic allows use of minimal doses of spinal local anesthetic with resultant rapid recovery and discharge (Table 47-4) and represents an alternative or complementary strategy to use of analgesic additives. Several clinical trials have determined minimally effective doses for ambulatory knee arthroscopy using this approach. For example, initial doses of spinal lidocaine 40 mg or mepivacaine 45 mg233,234 have been determined to be optimal doses for CSEA by previous dose–response studies. However, induction of a CSEA technique probably takes more time than conventional spinal anesthesia, and no current data are available to assess relative cost benefit of increased induction time versus decreased recovery time with CSEA. Thus final conclusions on the role of CSEA in a busy ambulatory surgery center remain to be determined.

Techniques and Equipment

The most widespread approach used in the literature is the needle-through-needle technique, in which an epidural needle is placed and then a spinal needle is inserted through it into the subarachnoid space. The spinal injection is made, and subsequently the spinal needle is removed from the epidural needle and an epidural catheter placed. A number of commercial kits are available. The simplest version is a Tuohy needle (or equivalent) through which a long, small-diameter spinal needle (24-30 gauge) is passed. Epidural needles with a "back hole" are also available, configured to allow placement of the spinal needle through a separate conduit to avoid bending of the spinal needle tip as it exits the curved epidural needle tip (Fig. 47-15). Recent studies suggest that use of a back-hole needle may offer advantages over a conventional needle-through-needle technique. A randomized trial in parturients observed decreased incidence of paresthesia (14% vs 42%) and failure to obtain CSF on the first attempt (8% vs 28%) with the back-hole needle.235 The separate conduit for the spinal needle may also reduce risk of toxicity from metal fragments caused by needle friction.236 Metal fragments have been proposed as a cause of aseptic meningitis, following the observation of notches in epidural needle tips.237 However, subsequent evaluations using atomic absorption spectrography and photomicrography did not demonstrate metal fragments, even after up to 5 spinal needle passes, and suggest that the notches were caused by malleability of the metal.238

As an alternative to the needle-through-needle technique, the double-segment method also offers the ability to place the epidural catheter and administer a test dose before placing the spinal block. Typically, the epidural and spinal blocks are performed separately at different interspaces. By first introducing the catheter, there exists the potential risk of damaging the catheter with the spinal needle. Furthermore, creating 2 separate cutaneous punctures could lead to increased incidence of adverse events, including backache, headache, infection, and hematoma.239 However, the double-segment technique has been shown to elicit fewer paresthesias as compared with a needle-through-needle approach.240 However, one study demonstrated greater acceptance by surgical patients of the needle-through-needle over the double-segment technique (85% vs 67%).241 That same study also showed a significantly longer time to perform the double-segment technique without decreasing the failure rate of spinal anesthesia, although other studies suggest a higher failure rate with the needle-through-needle technique.241

Potential Complications

Failure of Spinal Anesthesia

The combined technique is associated with a higher failure rate of spinal anesthesia than conventional spinal anesthesia. There are a number of reasons for failure: (1) Smaller-diameter spinal needles with long lengths are typically used. These needles lead to slower return of CSF and a greater resistance to injection. (2) Because the epidural needle has penetrated the tissue planes, there is little to anchor the spinal needle in place. Although a Luer lock apparatus is available, it locks at a fixed needle length, which can prevent the spinal needle from reaching or traversing the meninges.33 (3) Any deviation from midline can lead to missing the dura altogether (Fig. 47-16). (4) If loss-of-resistance technique uses saline, a false return of saline in the spinal needle rather than CSF can occur. Kopacz and Bainton7 recommend the hanging drop method within the epidural needle to aid identification of dural puncture in this situation. Negative pressure from tenting the dura with the spinal needle will cause an inward movement of the drop of fluid followed by return of CSF. (5) Finally, patient positioning and duration between spinal injection and completion of epidural catheter placement can change the characteristics of the spinal block.

Failure of Epidural Anesthesia

There are no controlled randomized prospective studies addressing the failure of epidural anesthesia or analgesia with the combined technique.28 The incidence of failure is unlikely to be higher with the combined technique, given the identical approach for standard epidural catheter placement. However, the difficulty in early testing with a needle-through-needle technique may lead to late recognition of a misplaced catheter. Prior injection of spinal anesthetic precludes testing the epidural catheter for intrathecal placement, and epidural injection of a test dose can lead to increased height of spinal block (see Intrathecal Effects of Epidural Agent).244 Reliability of detecting an intravascular test dose using 15 μg of epinephrine remains intact in healthy individuals using heart rate and systolic blood pressure criteria, although the developing spinal block will reduce the magnitude of hemodynamic response to epinephrine.216

Intrathecal Effects of Epidural Agent

The intrathecal effects of epidurally administered drugs can occur through migration of the epidural catheter through the meningeal puncture, leakage of epidural anesthetic through the meningeal hole, and pressure from the epidural injection displacing CSF and the local anesthetic suspended within it. The likelihood of passing an epidural catheter through a meningeal hole is very small, provided a 24-gauge or smaller spinal needle is used. This has been demonstrated using in vitro models and in vivo epiduroscopy.245 However, intrathecal catheter placement is possible if the epidural needle (17-18 gauge) initially rented the meninges. Migration later in the anesthetic course appears no more likely than with conventional epidural techniques.

Although Kamiya et al246 demonstrated that lidocaine concentrations in CSF are not different between epidural administration with or without prior meningeal puncture, significant leakage of epidural agents can occur through large meningeal rents, such as with a "wet tap."247,248

Pressure effect is the observation that increasing epidural volume can "squeeze" the CSF compartment and thus raise the cephalad spread of spinal drugs. A recent myelographic evaluation demonstrated that the subarachnoid space's diameter decreased to 25% after 10 mL of normal saline was injected through an epidural catheter.249 The ability to increase dermatomal spread by epidural volume appears to be time dependent. Sensory block extension can be significant (3-4 dermatomes) if epidural saline or air is injected soon after or before bupivacaine spinal anesthesia.243,250 This block enhancement may be clinically significant, as a clinical trial reported that CSEA required 20% less local anesthetic than single-shot spinal anesthesia.251 However, if delayed until 2-segment regression has begun, there is no increase in sensory blockade level252; in fact, it can even result in shorter duration of anesthesia.253

• Auroy Y, Benhamou D, Bargues L, et al. Major complications of regional anesthesia in France: the SOS Regional Anesthesia Hotline Service. Anesthesiology. 2002;97:1274-1280.
• Carpenter RL, Caplan RA, Brown DL, et al. Incidence and risk factors for side effects of spinal anesthesia. Anesthesiology. 1992;76:906-916.
• Curatolo M, Scaramozzino P, Venuti FS, et al. Factors associated with hypotension and bradycardia after epidural blockade. Anesth Analg. 1996;83(5):1033-1040.
• Hebl JR. The importance and implications of aseptic techniques during regional anesthesia. Reg Anesth Pain Med. 2006;31(4):311-323.
• Horlocker TT, Wedel DJ, Rowlingson JC, et al. Regional anesthesia in the patient receiving antithrombotic or thrombolytic therapy: American Society of Regional Anesthesia and Pain Medicine Evidence-Based Guidelines (Third Edition). Reg Anesth Pain Med. 2010;35(1):64-101.
• Liu SS, McDonald SB. Current issues in spinal anesthesia. Anesthesiology. 2001;94:888-906.
• Neal JM, Bernards CM, Butterworth JF, et al. ASRA practice advisory on local anesthetic systemic toxicity. Reg Anesth Pain Med. 2010;35(2):152-161.
• Neal JM, Bernards CM, Hadzic A, et al. ASRA Practice Advisory on Neurologic Complications in Regional Anesthesia and Pain Medicine. Reg Anesth Pain Med. 2008;33(5):404-415.