The spine consists of 33 vertebrae (7 cervical, 12 thoracic, 5 lumbar, 5 fused sacral, and 4 fused coccygeal; Fig. 34-1). With the exception of C1, the cervical, thoracic, and lumbar vertebrae consist of a body anteriorly, two pedicles that project posteriorly from the body, and two laminae that connect the pedicles (Fig. 34-2). These structures form the vertebral canal, which contains the spinal cord, spinal nerves, and epidural space. The laminae give rise to the transverse processes that project laterally and the spinous process that projects posteriorly. These bony projections serve as sites for muscle and ligament attachments. The pedicles contain a superior and inferior vertebral notch through which the spinal nerves exit the vertebral canal. The superior and inferior articular processes arise at the junction of the lamina and pedicles and form joints with the adjoining vertebrae. The first cervical vertebra (“atlas”) differs from this typical structure in that it does not have a body or a spinous process.

The five sacral vertebrae are fused together to form the wedge-shaped sacrum, which connects the spine with the iliac wings of the pelvis (Fig. 34-1). The fifth sacral vertebra is not fused posteriorly, giving rise to a variably shaped opening known as the sacral hiatus. Occasionally, other sacral vertebrae do not fuse posteriorly, giving rise to a much larger sacral hiatus. The sacral cornua
are bony prominences on either side of the hiatus and aid in identifying it. The sacral hiatus provides an opening into the sacral canal, which is the caudal termination of the epidural space. The four rudimentary coccygeal vertebrae are fused together to form the coccyx, a narrow triangular bone that abuts the sacral hiatus and can be helpful in identifying it. The tip of the coccyx can often be palpated in the proximal gluteal cleft, and by running one’s finger cephalad along its smooth surface, the sacral cornu can be identified as the first bony prominence encountered.

Identifying individual vertebrae is important for correctly locating the desired interspace for epidural and spinal blockade. The C7 and T1 spinous processes are the most prominent and are easily palpated while running the hand down the back of the neck. The twelfth thoracic vertebra can be identified by palpating the twelfth rib and tracing it back to its attachment to T12. A line drawn between the iliac crests crosses the body of L5 or the L4 to L5 interspace.

The vertebral bodies are stabilized by five ligaments that increase in size between the cervical and lumbar vertebrae (Fig. 34-2). From the sacrum to T7, the supraspinous ligament runs between the tips of the spinous processes. Above T7, this ligament continues as the ligamentum nuchae and attaches to the occipital protuberance at the base of the skull. The interspinous ligament attaches between the spinous processes and blends posteriorly with the supraspinous ligament and anteriorly with the ligamentum flavum. The ligamentum flavum is a tough, wedge-shaped ligament composed of elastin. It consists of right and left portions that span adjacent vertebral laminae and fuse in the midline to varying degrees.⁷,⁸ The ligamentum flavum is thickest in the midline, measuring 3 to 5 mm at the L2 to L3 interspace of adults. The distance between the ligamentum flavum and the spinal meninges is also greater in the midline, measuring 4 to 6 mm at the L2 to L3 interspace.⁹ As a result, midline insertion of an epidural needle is least likely to result in unintended meningeal puncture. The anterior and posterior longitudinal ligaments run along the anterior and posterior surfaces of the vertebral bodies.

The epidural space is the space that lies between the spinal meninges and the sides of the vertebral canal (Fig. 34-3). It is bounded cranially by the foramen magnum, caudally by the sacrococcygeal ligament covering the sacral hiatus, anteriorly by the posterior longitudinal ligament, laterally by the vertebral pedicles, and posteriorly by both the ligamentum flavum and the vertebral lamina. The epidural space is not a closed space but communicates with the paravertebral space by way of the intervertebral foramina.¹⁰ The epidural space is shallowest anteriorly where the dura may in some places fuse with the posterior longitudinal ligament. The space is deepest posteriorly, although the depth varies because the space is intermittently obliterated by contact between the dura mater and the ligamentum flavum or vertebral lamina. Contact between the dura mater and the pedicles also interrupts the epidural space laterally. Thus, the epidural space is composed of a series of discontinuous compartments that become continuous when the potential space separating the compartments is opened up by injection of air or liquid. A rich network of valveless veins (Batson plexus) courses through the anterior and lateral portions of the epidural space and only very rarely in the posterior epidural space (Fig. 34-2).¹¹ The epidural veins anastomose freely with extradural veins, including the pelvic veins, azygous system, and intracranial veins. The epidural space also contains lymphatics and segmental arteries running between the aorta and the spinal cord.

The spinal meninges consist of three protective membranes (dura mater, arachnoid mater, and pia mater), which are continuous with the cranial meninges (Fig. 34-4).

CSF is a complex solution made up of 99% water and containing an array of molecules including electrolytes, proteins, glucose,
neurotransmitters, neurotransmitter metabolites, cyclic nucleotides, amino acids, among many others. The CSF volume is approximately 100 to 160 mL in adult humans and it is produced at the rate of 20 to 25 mL/hr. Consequently, the entire CSF volume is replaced roughly every 6 hours.

According to the traditional view of CSF hydrodynamics, CSF is produced by plasma ultrafiltration in the choroid plexus within the cerebral ventricles. It then “flows” or “circulates” unidirectionally through the subarachnoid space to finally be absorbed by the arachnoid villi in the superior sagittal sinus and/or into lymphatics via perineural sheaths of cranial and spinal nerves. Several recent investigations, however, contradict this widely held view. Bulat et al.²⁴ hypothesized that CSF is formed by water filtration across arterial capillary walls in the CNS. The filtration of water from arterial capillaries generates an osmotic counterpressure that drives water reabsorption into cerebral microvessels (i.e., venous capillaries and postcapillary venules). Using radiolabeled water, they demonstrated that CSF does not flow through subarachnoid spaces and that the bulk removal of CSF occurred primarily via absorption into cerebral microvessels. Klarica et al.²⁵ further demonstrated that occlusion of the aqueduct of Sylvius does not alter isolated cerebral ventricle pressures, suggesting that the formation and absorption of CSF are in balance with one another.

In addition, the development of cine-magnetic resonance imaging and cine-computed tomography techniques have also shown that CSF does not circulate, rather it oscillates in the cephalocaudal axis with a frequency equal to the heart rate.²⁶,²⁷ CSF oscillates because cerebral expansion during systole displaces CSF caudally into the spinal canal and cerebral contraction during diastole causes the CSF displaced into the spinal canal to retreat back into the cranial vault. Net CSF movement is estimated at 0.04% per oscillation.

The clinical significance of this understanding of CSF motion is that CSF cannot be relied on to distribute drugs in the subarachnoid space. This is of little importance in single-shot spinal anesthesia because the kinetic energy of the injection and the baricity of the solution serve to distribute the drug. However, the lack of significant net CSF motion explains why drug distribution during the very slow infusions used for chronic intrathecal analgesia results in very limited drug distribution.²⁸

In the first-trimester fetus, the spinal cord extends from the foramen magnum to the end of the sacrum. Thereafter, the vertebral column lengthens more than the spinal cord so that at birth the spinal cord ends at about the level of the third lumbar vertebra. In the adult, the caudad tip of the spinal cord typically lies at the level of the first lumbar vertebra. However, in 30% of individuals the spinal cord may end at T12, while in 10% it may extend to L3 (Fig. 34-1).²⁹ A sacral spinal cord has been reported in an adult.²⁹ Flexion of the vertebral column causes the tip of the spinal cord to move slightly cephalad.

The spinal cord gives rise to 31 pairs of spinal nerves, each composed of an anterior motor root and a posterior sensory root. The nerve roots are in turn composed of multiple rootlets. The portion of the spinal cord that gives rise to all of the rootlets of a single spinal nerve is called a cord segment. The skin area innervated by a given spinal nerve and its corresponding cord segment is called a dermatome (Fig. 34-5). The intermediolateral gray matter of the T1 through L2 spinal cord segments contains the cell bodies of the preganglionic sympathetic neurons. These sympathetic neurons run with the corresponding spinal nerve to a point
just beyond the intervertebral foramen where they exit to join the sympathetic chain ganglia.

The spinal nerves and their corresponding cord segments are named for the intervertebral foramen through which they run. In the cervical region, the spinal nerves are named for the vertebra forming the caudad half of the intervertebral foramen; for example, C4 emerges through an intervertebral foramen formed by C3 and C4. In the thoracic and lumbar regions, the nerve roots are named for the vertebrae forming the cephalad half of the intervertebral foramen; for example, L4 emerges through an intervertebral foramen formed by L4 and L5. Since the spinal cord usually ends between L1 and L2, the thoracic, lumbar, and sacral nerve roots run increasingly longer distances in the subarachnoid space to get from their spinal cord segment of origin to the intervertebral foramen through which they exit. Those nerves that extend beyond the end of the spinal cord to their exit site are collectively known as the cauda equina (Fig. 34-1).

Spinal and epidural needles are classified by the design of their tips (Fig. 34-6). The Whitacre, Eldor, Marx, and Sprotte spinal needles have a “pencil-point” tip with one or two (Eldor) aperture(s) on the side of the shaft proximal to the tip. The Greene, Atraucan, and Quincke needles have beveled tips with cutting edges. The pencil-point needles require more force to insert than the bevel-tip needles but provide a better tactile “feel” of the various tissues encountered as the needle is inserted. In addition, the bevel has been shown to cause the needle to be deflected from the intended path as it passes through tissues while the pencil-point needles are not deflected.³⁰ Epidural needles have a larger diameter than spinal needles to facilitate the injection of fluid or air when using the “loss of resistance” technique to identify the epidural space. In addition, the larger diameter allows for easier insertion of catheters into the epidural space. The Tuohy epidural needle has a curved tip to help control the direction that the catheter moves in the epidural space. The Hustead needle tip is also curved, although somewhat less than the Tuohy needle. The Crawford needle tip is straight, making it less suitable for catheter insertion.

The outside diameter of both epidural and spinal needles is used to determine their gauge. Larger gauge (i.e., smaller diameter) spinal needles are less likely to cause postdural puncture headaches (PDPH), but are more readily deflected than smaller gauge needles. Epidural needles are typically sized 16G to 19G and spinal needles 22G to 29G. Spinal needles smaller than 22G are often easier to insert if an introducer needle is used. The introducer is inserted into the interspinous ligament in the intended direction of the spinal needle and the spinal needle is then inserted through the shaft of the introducer. The introducer prevents the spinal needle from being deflected or bent as it passes through the interspinous ligament.³⁰ Needles of the same outside diameter may have different inside diameters. This is important because inside diameter determines how large a catheter can be inserted through the needle and determines how rapidly CSF will appear at the needle hub during spinal needle insertion. All spinal and epidural needles come with a tight fitting stylet. The stylet prevents the needle from being plugged with skin or fat and, importantly, prevents dragging skin into the epidural or subarachnoid spaces, where the skin may grow and form dermoid tumors.

If the patient desires, light sedation is appropriate before placement of spinal or epidural block. Generally, the patient should not be heavily sedated because successful spinal and epidural anesthesia requires patient participation to maintain good position, evaluate block height, and to enable communication with the anesthesiologist should a paresthesia occur when the needle contacts neural elements. In addition, patient cooperation is
required to properly evaluate an epidural test dose; sedation with as little as 1.5 mg of midazolam plus 75 μg of fentanyl has been shown to reduce the reliability of patient reports of subjective symptoms of intravenous (IV) local anesthetic injection.³¹ Once the block is placed and adequate block height assured, the patient can be sedated as deemed appropriate.

Inserting a catheter into the subarachnoid space increases the utility of spinal anesthesia by permitting repeated drug administration as often as necessary to extend the level or duration of spinal block. A common and reasonable recommendation for subsequent dosing or “topping up” of continuous spinal blocks is to administer half the original dose of local anesthetic when the block has reached two thirds of its expected duration.

The technique is similar to that described for “single-shot” spinal anesthesia except that a needle large enough to accommodate
the desired catheter must be used. After inserting the needle and obtaining free-flowing CSF, the catheter is simply threaded into the subarachnoid space a distance of 2 to 3 cm. It is often easier to insert the catheter if it is directed cephalad or caudad instead of laterally. If the catheter does not easily pass beyond the needle tip, rotating the needle 180 degrees may be helpful or another interspace may be used. Once the catheter is advanced beyond the needle tip, the catheter should not be withdrawn back into the needle shaft due to the risk of shearing the catheter off into the subarachnoid space.

A variety of catheters and needles are available for continuous spinal anesthesia. Commonly, 18G epidural needles and 20G catheters are used. However, needles and catheters this size carry a higher risk of PDPH, especially in young patients. Because of this risk, smaller needle and catheter combinations have been developed with catheters ranging in size from 24G to 32G. Although smaller catheters decrease the risk of PDPH, they have also been associated with multiple reports of neurologic injury, specifically, cauda equina syndrome (see “Complications”). For this reason, the United States Food and Drug Administration has advised against using any catheter smaller than 24G for continuous spinal anesthesia.

For the novice, correct placement of an epidural needle can be technically more challenging than spinal needle placement because there is less room for error. However, with experience, epidural needle placement is often easier than spinal needle placement because the larger gauge needles used for epidural anesthesia are less likely to be deflected from their intended path and they produce much better tactile feel of the interspinous and flaval ligaments. In addition, the loss of resistance technique provides a much clearer end point when entering the epidural space than does the subtle “pop” of a spinal needle piercing the dura mater.

Patient preparation, positioning, monitors, and needle approaches for epidural anesthesia are the same as for spinal anesthesia. Unlike spinal anesthesia, epidural anesthesia may be performed at any intervertebral space. However, at vertebral levels above the termination of the spinal cord, the epidural needle may accidentally puncture the spinal meninges and damage the underlying spinal cord. To prevent accidental meningeal puncture, the anesthesiologist must learn to identify the interspinous ligaments and the ligamentum flavum by their feel. In addition, epidural needles must be advanced slowly and, most importantly, under control.

After proper positioning, sterile skin preparation, and draping, the desired interspace is identified and a local anesthetic skin wheal is raised at the point of needle insertion. Because epidural needles are relatively blunt, it is sometimes helpful to pierce the skin with a ≥18G hypodermic needle before inserting the epidural needle. For epidural anesthesia using the midline approach, the epidural needle is inserted through the subcutaneous tissue and into the interspinous ligament. The interspinous ligament has a characteristic “gritty” feel, much like inserting a needle into a bag of sand. This is especially true of younger patients. If the interspinous ligament is not clearly identified, then one should be suspicious that the needle is not in the midline. After engaging the interspinous ligament, the needle is advanced slowly through it until an increase in resistance is felt. This increased resistance represents the ligamentum flavum.

The epidural needle must now traverse the ligamentum flavum and stop within the epidural space before puncturing the spinal meninges. Numerous techniques for identifying the epidural space have been used successfully; however, the loss of resistance to fluid has the advantage of simplicity, reliability, and, most importantly, a higher success rate when compared to the use of air for loss of resistance.³² In addition, use of fluid instead of air for loss of resistance decreases the risk of PDPH in the event of accidental meningeal puncture.³³

A glass syringe or a specially designed low-resistance plastic syringe is filled with 2 to 3 mL of saline and a small (0.1 to 0.3 mL) air bubble. The syringe is attached to the epidural needle and the plunger is pressed until the air bubble is visibly compressed. If the needle tip is properly embedded within the ligamentum flavum, it should be possible to compress the air bubble without injecting fluid. In this way the air bubble serves as a gauge of the appropriate amount of pressure to exert on the syringe plunger. If the air bubble cannot be compressed without injecting fluid, then the needle tip is most likely not in the ligamentum flavum. In this case, the needle tip may still be in the interspinous ligament, or it may be off the midline in the paraspinous muscles. To differentiate between these possibilities, one can carefully advance the needle and syringe a few millimeters in an effort to engage the ligamentum flavum. If it is still not possible to compress the air bubble, withdraw the needle into the subcutaneous tissue and reinsert it.

Once the ligamentum flavum is identified, the needle is slowly advanced with the nondominant hand while the dominant hand maintains constant pressure on the syringe plunger (Fig. 34-8). Do not advance the needle with the hand compressing the plunger because this does not allow for adequate control of needle movement. As the needle tip enters the epidural space, there will be a sudden and dramatic loss of resistance as the saline is rapidly injected. Saline injection into the epidural space can be moderately painful and patients should be forewarned. If the needle is advancing obliquely through the ligamentum flavum, it is possible to enter into the paraspinous muscles instead of the epidural space. In this case the loss of resistance will be less dramatic. To
help verify that the needle has entered the epidural space, 0.5 mL of air can be drawn into the syringe and injected. There will be virtually no resistance to air injection in the epidural space, while in the paraspinous muscles air injection will encounter demonstrable resistance.

After entering the epidural space, stop advancing the needle. Because the dura mater abuts the ligamentum flavum in many places, the dura may now be tented over the needle tip and advancing the needle any farther than necessary heightens the risk of accidental meningeal puncture, that is, “wet tap.” When the syringe is disconnected from the needle, it is common to have a small amount of fluid flow from the needle hub. This is usually the saline flowing back out of the epidural space but it could be CSF if the needle accidentally entered the subarachnoid space. CSF can often be distinguished by the fact that CSF will usually flow out in a volume greatly exceeding that used for the loss of resistance, CSF will be warm compared with saline, and CSF will test positive for glucose.

If a single-shot technique is to be used, then a local anesthetic test dose should be administered to help rule out undetected subarachnoid or IV needle placement. After a negative test dose, the desired volume of local anesthetic should be administered in small increments (e.g., 5 mL) at a rate of 0.5 to 1 mL/sec. Slow, incremental injection decreases the risk of pain during injection and allows detection of adverse reactions to accidental IV or subarachnoid placement before the entire dose is administered.

Use of a catheter for epidural anesthesia affords much greater flexibility than the single-shot technique because the catheter can be used to prolong a block that is too short, to extend a block that is too low, or to provide postoperative analgesia. On the downside, catheters may migrate into an epidural vein, into the subarachnoid space, or out an intervertebral foramen. Catheter use is also more likely to result in unilateral epidural block, a clinical fact shown to result from catheter tips that end up in the anterior epidural space or migrate out an intervertebral foramina.¹⁸,³⁴ An ever-changing selection of epidural catheters is commercially available. They differ in diameter, stiffness, location of injection holes, presence or absence of a stylet, construction material, and the like. Whichever catheter is chosen, it is important to verify that it passes easily through the epidural needle before the needle is placed in the epidural space. Epidural catheters are usually inserted through either Tuohy or Hustead needles because their curved tips help direct the catheter away from the dura mater. The needle bevel should be directed either cephalad or caudad, although the direction of the bevel does not guarantee that the catheter will travel in that direction. The catheter will typically encounter resistance as it reaches the curve at the tip of the needle, but steady pressure will usually result in passage into the epidural space. If the catheter will not pass beyond the needle tip, it is possible that the needle opening is not completely in the epidural space or that some structure in the epidural space is preventing catheter insertion (e.g., epidural fat). In this instance, the needle can be carefully advanced 1 to 2 mm more or rotated 180 degrees and the catheter reinserted. Although either of these maneuvers may result in successful catheter placement, they also increase the risk of accidental meningeal puncture. Alternatively, the procedure can be repeated at another interspace or with a different needle approach, for example, paramedian. Occasionally a catheter will advance only a short distance past the needle tip. This raises the possibility that the needle tip is not in the epidural space and needs to be repositioned. In this case, the catheter should not be withdrawn back into the epidural needle because of the risk that the catheter tip will be sheared off by the bevel’s sharp edge. Rather, the needle and catheter should be pulled out in tandem and the procedure repeated. An alternative explanation for the inability to thread an epidural catheter is that the tip of the epidural needle was bent during bony contact and now partially occludes the needle lumen.

The catheter should be advanced only 3 to 5 cm into the epidural space. Placing a longer length of catheter in the epidural space increases the risk that it will form a knot,³⁵,³⁶ enter an epidural vein, puncture the spinal meninges, exit an intervertebral foramen, wrap around a nerve root, or wind up in some other disadvantageous location. Once the catheter is appropriately positioned in the epidural space, the needle is slowly withdrawn with one hand as the catheter is stabilized with the other. After the needle is removed, the length of the catheter in the epidural space is confirmed by subtracting the distance between the skin and the epidural space from the length of the catheter below the skin. Documenting this distance is important when trying to determine if catheters used in the postoperative period have been dislodged.

An epidural test dose must be administered through the catheter to test for IV or subarachnoid placement before incrementally delivering the entire epidural drug dose. In addition, because of the risk of undetected IV or subarachnoid migration of the catheter over time, additional test doses must be administered before each top-up dose is given through the catheter. As with continuous spinal anesthesia, a reasonable guideline for top-up doses is to administer half the initial local anesthetic dose at an interval equal to two thirds the expected duration of the block.

The epidural test dose is designed to identify epidural needles or catheters that have entered an epidural vein or the subarachnoid space. Failure to perform the test may result in IV injection of toxic doses of local anesthetic or total spinal block. Aspirating the catheter or needle to check for blood or CSF is helpful if positive, but the incidence of false-negative aspirations is too high to rely on this technique alone.³⁷

The most common test dose is 3 mL of local anesthetic containing 5 μg/mL of epinephrine (1:200,000). The dose of local anesthetic should be sufficient that subarachnoid injection will result in clear evidence of spinal anesthesia. Intravenous injection of this dose of epinephrine typically produces an average 30 beats per minute heart rate increase between 20 and 40 seconds after injection.³⁸,³⁹ Heart rate increases may not be as evident in some patients taking beta-blocking drugs; reflex bradycardia often occurs in these patients.³⁸,⁴⁰ In beta-blocked patients, a systolic blood pressure increase of ≥20 mm Hg may be a more reliable indicator of IV injection.³⁸,⁴⁰

Importantly, the sensitivity of the standard 15-μg epinephrine test dose has been shown to be markedly diminished by pre-existing high thoracic epidural anesthesia and/or concurrent general anesthesia.⁴¹ Larger epinephrine doses may be effective at detecting IV injection in these settings, but that has not been shown experimentally.

Isoproterenol has also been used to detect intravascular injection.⁴² In addition, air injection combined with a precordial Doppler to detect the characteristic murmur has been used successfully to test for IV placement of epidural catheters.³⁷ These techniques have been developed for the use in laboring women in whom the sensitivity of epinephrine as a test dose is disturbingly low because maternal heart rate increases during contractions are
often as large as those produced by epinephrine.⁴³ The clinical indications for these alternative tests of IV injection await additional larger studies.

Combined spinal–epidural anesthesia (CSEA) is a useful technique by which a spinal block and an epidural catheter are placed simultaneously. This technique is popular because it combines the rapid onset, dense block of spinal anesthesia with the flexibility afforded by an epidural catheter. There are special epidural needles with a separate lumen to accommodate a spinal needle available for CSEA (Fig. 34-6). However, the technique is easily performed by first placing a standard epidural needle in the epidural space and then inserting an appropriately sized spinal needle through the shaft of the epidural needle and into the subarachnoid space. The desired local anesthetic is injected into the subarachnoid space, the spinal needle is removed, and a catheter placed in the epidural space via the epidural needle. The catheter can then be used to extend the height or duration of intraoperative block or can be used to provide postoperative epidural analgesia.

An interesting pharmacologic aspect of CSEA is the observation that after the peak spinal block height is established, both saline and local anesthetic injected into the epidural space are effective at pushing the block level higher.⁴⁴,⁴⁵,⁴⁶ This observation has been interpreted to indicate that the mechanism by which the epidural “top-up” increases block height is by a volume effect (i.e., compression of the spinal meninges forcing CSF cephalad) as well as a local anesthetic effect.

A potential risk of this technique is that the meningeal hole made by the spinal needle may allow dangerously high concentrations of subsequently administered epidural drugs to reach the subarachnoid space. Anecdotal case reports and in vitro animal studies suggest that this may be a legitimate concern.⁴³,⁴⁷,⁴⁸,⁴⁹ Although CSEA is advantageous in some circumstances, additional prospective studies are necessary to identify the relative risks and limitations of the technique.