The Spinal Cord, Spinal Canal, and Meninges

The cephalad aspect of the spinal cord is continuous with the brainstem through the foramen magnum. The spinal cord most often terminates as the conus medullaris at the level of the lower border of the first lumbar vertebral body. The conus medullaris is attached to the coccyx by means of a neural-fibrous band called the filum terminale, which is surrounded by the nerves of the lower lumbar and sacral roots, known as the cauda equina. Within the bony vertebral column are three membranes: the pia mater, the arachnoid mater, and the dura mater. The pia mater is a highly vascular membrane that closely invests the spinal cord. The arachnoid mater is a delicate, nonvascular membrane closely adherent to the third and outermost layer, the dura. The subarachnoid space, located between the pia mater and arachnoid mater, contains (1) cerebrospinal fluid (CSF), (2) spinal nerves, (3) a trabecular network between the two membranes, (4) blood vessels that supply the spinal cord, and (5) lateral extensions of the pia mater—the dentate ligaments. The dura mater is a membrane composed of collagen that encapsulates the spinal cord, the deeper meningeal layers, and the subarachnoid space. This layer forms a connective tissue sheath along the vertical axis of the central nervous system (CNS) that is contiguous with connective tissue covering the lateral extension of spinal nerve roots as they exit the intervertebral foramina. The interface between the dural and arachnoid layers has been described as a potential space capable of expansion subsequent to mechanical trauma. Unintentional injection of local anesthetic into this subdural space may explain some cases of failed spinal anesthesia. It may also explain the rare, slow-to-develop cases of high spinal anesthesia after the inadvertent subdural injection of larger volumes of local anesthetic intended for epidural administration.¹ Although the spinal cord ends at the level of the bodies of L1 and L2 in most patients, the subarachnoid space and cauda equina continue to the S2 level (Figure 12-1).

The epidural space is located external to the sac of the dura mater and contains loose connective tissue, adipose tissue, lymphatics, spinal nerve roots, and the internal vertebral venous plexus (Batson’s plexus) (Figure 12-2). This space is bound by the posterior longitudinal ligament anteriorly, the ligamentum flavum and the periosteum of the lamina posteriorly, the pedicles of the vertebrae, and the intervertebral foramina with their contents laterally. The epidural space is closed at the foramen magnum where the spinal dura attaches to the dura of the cranium and at the sacral hiatus by the sacrococcygeal ligament. Frequently, anatomic references will illustrate neuraxial anatomy by way of sagittal and/or transverse cross section. This may result in the erroneous impression that the epidural space is a continuous columnar entity that envelops the dural sac at all points about its perimeter. Investigations using cryomicrotome sections and three-dimensional reconstruction of radiologic data verify that the epidural space is in fact discontinuous along the vertical and lateral axes of the spinal canal. It varies in anteroposterior thickness according to dermatomal distribution, being widest at the level of lumbar vertebrae and thinnest in the cervical region.^2,3 Epidu­roscopy and epidurography suggest the presence of a dorsal median connective tissue band in some individuals.⁴ Anatomic dissection and computed tomographic epidurography have also suggested the presence of epidural space septa. This band (or these septa) may provide an explanation for unilateral or incomplete epidural anesthesia.⁵ However, some investigators have suggested that the dorsal median band is an artifact of epidural space distention or an anatomic manifestation of the previously unappreciated epidural space segmentation.³

The ligamentum flavum lies posterior to the epidural space (Figure 12-3). The lamina, the spinous processes of the vertebral bodies, and the interspinous ligaments lie posterior to the ligamentum flavum. Posterior to these structures are the supraspinous ligament (which extends from the external occipital protuberance to the coccyx), subcutaneous tissue, and skin. Historically, some have described the ligamentum flavum as a single ligament. In actuality, however, it is composed of two curvilinear ligaments that join in the middle and form an acute angle with a ventral opening.^3,6 Much like the epidural space, the ligamentum flavum is not uniform from skull to sacrum; indeed, it is not uniform even within a single intervertebral space. The thickness of the ligamentum flavum varies with vertebral level, body mass index, and age, as does the distance between the skin and the epidural space (Table 12-1).^7,8

Monitoring

The American Society of Anesthesiologists (ASA) has published guidelines for administration of neuraxial anesthesia in obstetric patients (see Appendix A). Among other things these recommendations address (1) the required presence of qualified anesthesia and obstetric care providers, (2) immediate availability of resuscitation medication and equipment (Box 12-1), (3) mandatory pre-procedural intravenous access, and (4) employment and documentation of maternal vital signs and fetal heart rate (FHR) monitoring.

Neither the ASA nor the American College of Obstetricians and Gynecologists (ACOG) provides a specific recommendation as to whether continuous FHR monitoring is necessary during performance of neuraxial anesthesia procedures. The ASA Task Force on Obstetric Anesthesia¹⁶ has stated:

The anesthesia provider cannot predict when hypotension will occur during the administration of neuraxial anesthesia. In addition, there is concern that intrathecal administration of an opioid is associated with a higher incidence of nonreassuring FHR patterns than other neuraxial techniques (see Chapter 23).¹⁷ Thus, we believe that continuous electronic FHR monitoring should be performed both during (if possible) and after the administration of neuraxial analgesia in all laboring women. In some cases, the mother’s position or maternal obesity precludes the use of an external Doppler device to monitor the FHR. In such cases (especially when there is concern regarding fetal well-being), it is helpful for the obstetric provider to place a fetal scalp electrode to monitor the FHR.

Prior to initiation of neuraxial anesthesia, a pre-procedural verification process (i.e., “time-out”) is instituted as part of compliance with national patient safety recommendations from hospital accreditation organizations.¹⁸ The participation of the patient, the anesthesia care provider, and a third party such as a member of the nursing staff may lead to the discovery of concerns that should be addressed before the initiation of neuraxial anesthesia. The risk of neuraxial procedures relates to (1) the physical instrumentation of the spinal axis and (2) physiologic changes associated with medication administration via this anatomic route. Contraindications to needle or catheter placement include patient refusal or inability to cooperate, ongoing bleeding diathesis, infection either at the site of intended intervention or untreated systemic blood-borne illness, and increased intracranial pressure predisposing to cerebral herniation. Contraindications to injecting local anesthetics via the epidural or spinal route include severe hypovolemia and allergy to local anesthetics. It would be axiomatic that patient refusal presents an absolute contraindication to an elective procedure. A thorough preoperative assessment of current fetal well-being, maternal volume status, intrapartum systemic opioid use, antibiotic administration for ongoing chorioamnionitis or other infectious process, and a brief reiteration of known maternal disease states, including allergies, will readily identify most of the major concerns that would render neuraxial anesthesia potentially hazardous. The anesthesia provider should weigh the risks and benefits of neuraxial anesthesia for each patient.

Informed consent should include a frank discussion about anesthetic procedures and risks. Surveys of postpartum women have demonstrated that most parturients want to know the possible complications of epidural analgesia, even those that are rare.^19,20 It is best to relay this information before the onset of labor (e.g., during antenatal classes),²¹ or early in the intrapartum period, although doing so is not always feasible. Some anesthesia providers fear that distressed, desperate, or sedated parturients may not understand the discussion of anesthetic procedures. However, adequacy of consent can be demonstrated not only by documentation of information provided to the patient but also by the lack of patient objection to a procedure and the cooperation provided by the patient during the procedure.²² In a survey of North American anesthesiologists, Brull et al.²³ found that a majority of clinicians did not disclose the most severe risks of central neuraxial anesthesia to their patients. Additionally, most anesthesiologists cited inaccurate incidences of these complications. We do not find it difficult to explain the procedure and the risks of neuraxial analgesia to a laboring woman. The preanesthetic evaluation allows the physician to communicate a sense of concern and to demonstrate a commitment to the patient’s care. Most laboring women understand the need for informed consent, and they appreciate the opportunity to participate in decisions about their care.

Management of the pregnant patient occurs in a unique clinical care environment in which the presence of the patient’s spouse or family member must be addressed. Most often, the dictates of local institutions will establish whether a partner’s presence during neuraxial labor analgesia is acceptable. However, hospital policy may be vague enough that the discretion lies in the hands of the anesthesia provider. Intuition may suggest that a partner who remains present during the conduct of neuraxial analgesia may help alleviate the patient’s ongoing anxiety regarding the procedure. Conversely, the partner may be so apprehensive or disruptive that the partner’s presence becomes counterproductive to the care of the patient. Orbach-Zinger et al.²⁴ randomized 84 nulliparous women to either presence or absence of their partner during labor epidural catheter placement. Interestingly, patient and partner anxiety, as measured by a validated anxiety questionnaire, were reduced when partners were absent during the procedure.

Patient Positioning

Pregnant women have an exaggerated lumbar lordosis, and it is more difficult for them to flex the lumbar spine. However, most pregnant women are young, and youth usually allows sufficient flexibility to facilitate the insertion of a needle into the epidural or subarachnoid space. Whether the block is initiated in the lateral or sitting position is a matter of provider and patient preference. Notable advantages of the lateral position include (1) orthostatic hypotension is less likely and (2) the position often facilitates continuous FHR monitoring during placement of the epidural catheter. Vincent and Chestnut²⁵ performed a study in which they observed that neither the sitting nor the lateral position was consistently superior with regard to patient comfort. However, pregnant women who preferred the left lateral decubitus position weighed less and had lower body mass indices than women who preferred the sitting position. The sitting position is likely associated with a higher incidence of orthostatic hypotension and syncope. However, the sitting position is preferred—and may be required—in obese parturients, in whom identification of the midline is usually significantly easier in the sitting position. Further, morbidly obese women may experience hypoxemia when placed in the lateral decubitus position.

One study demonstrated a greater reduction in maternal cardiac output with maximal lumbar flexion in the lateral decubitus position than in the sitting position during identification of the epidural space in laboring women.²⁶ The researchers speculated that maximal lumbar flexion in the lateral decubitus position results in concealed aortocaval compression. In contrast, they suggested that the uterus falls forward (and thus does not cause aortocaval compression) when the patient assumes the sitting flexed position. They recommended that “the tight fetal curl position be avoided,” especially when the patient assumes the lateral decubitus position for identification of the epidural or intrathecal space.

Aortocaval compression must be avoided at all times. The gravid uterus can occlude the inferior vena cava and aorta when the parturient assumes the supine position.^27–29 This position may cause maternal hypotension^30,31 and reduce uteroplacental perfusion,³² even in the absence of anesthesia. Increased venous tone in the lower extremities helps overcome partial occlusion of the inferior vena cava in unanesthetized pregnant women. If maternal hydration is inadequate and if aortocaval compression is not avoided, the onset of anesthesia-induced sympathetic blockade may result in decreased venous return, cardiac output, and uteroplacental perfusion.²⁹

Maternal position during placement of the epidural catheter does not seem to affect the incidence of unintentional dural puncture. However, adoption of the lateral recumbent head-down position for epidural catheter placement may reduce the incidence of epidural venous puncture.³³

When spinal or epidural anesthesia is performed with the patient in a lateral position, the patient’s back should lie at, and parallel to, the edge of the bed, for at least two reasons. First, the edge is the most firm section of the mattress. If the patient lies away from the edge of the bed, the patient’s weight will depress the mattress, and the anesthesia provider must work in a “downhill” direction. Second, this position allows anesthesia providers to keep their elbows flexed, facilitating control of fine hand and wrist muscle movements. The plane of the entire back should be perpendicular to the mattress. When asked to flex the lower back, patients typically roll the top shoulder forward, an action that rotates the spine (which is undesirable) but does not flex the lower back.

Similarly, patients positioned sitting should have their feet supported by a stool with the backs of their knees against the edge of the bed, a maneuver that helps position the patient’s back closer to the anesthesia provider. The shoulders should be relaxed symmetrically over the hips and buttocks. Beds in obstetric units often break at the foot, and the split in the mattress encourages the patient’s seat to slope downhill if she is straddling the mattress split; this position will cause spine rotation and may make the procedure more difficult.

When spinal anesthesia is performed, the patient’s posture relative to the baricity of the anesthetic solution should be considered, because it influences the extent of blockade, the latency of blockade, and the incidence of hypotension. The incidence, timing, and extent of hypotension in the period immediately after initiation of the block depend on the type of block (e.g., spinal, epidural, or combined spinal-epidural [CSE]), drug characteristics (e.g., baricity, concentration), patient position during the procedure, and patient position in the period following the procedure. For example, when spinal anesthesia is initiated with a hyperbaric solution for instrumental vaginal delivery, it often makes sense for the patient to be sitting to ensure the rapid onset of sacral anesthesia. Conversely, spinal anesthesia for cervical cerclage can be initiated with the patient in the steep lateral Trendelenburg position with a hypobaric anesthetic solution.

Posture has less influence on the spread of epidural anesthesia.^34–36 During epidural anesthesia, a unilateral block more likely results from the malposition of the catheter (or perhaps an anatomic barrier within the epidural space) than from patient position, particularly after a bolus injection. Norris and Dewan³⁴ observed that gravity did not augment the spread of anesthesia in patients receiving epidural anesthesia for cesarean delivery, and they concluded that posture does not need to be manipulated to ensure adequate bilateral epidural anesthesia. In at least two studies it was noted that the use of the sitting position is not necessary for the development of good sacral anesthesia when large volumes of epidural local anesthetic are given for cesarean delivery.^34,36 However, Reid and Thorburn³⁶ observed that use of the sitting position appeared to delay the spread of anesthesia to the midthoracic dermatomes. In comparison with the bolus administration of epidural local anesthetic, the extent of blockade may be more gravity dependent when the anesthetic is administered as a continuous infusion over a prolonged period.

Some anesthesiologists contend that maternal position after epidural catheter placement affects the efficacy of epidural analgesia, although this is a matter of some dispute. Beilin et al.³⁷ observed that the placement of the laboring woman in the supine position with a 30-degree leftward tilt was associated with better epidural analgesia than maintenance of the left lateral decubitus position. In contrast, Preston et al.³⁸ observed no difference in analgesia and a significantly higher incidence of fetal bradycardia with the supine wedged position than with the full lateral position.

Caudal anesthesia is used infrequently in modern obstetric anesthesia practice. However, there remain some circumstances in which a caudal technique is useful and/or advantageous. It is a good choice for the second stage of labor in selected patients in whom the lumbar epidural approach is hazardous or contraindicated (e.g., fusion or instrumentation of the lumbar spine). In most cases, caudal anesthesia can be successfully performed with the patient in a lateral decubitus position.

Aseptic Technique

There appears to be a great degree of practice variation among anesthesia providers with respect to aseptic technique during administration of neuraxial anesthesia.^39–41 This lack of consensus may result from an underappreciation of the gravity of infectious complications related to neuraxial anesthesia.^42–44 Indeed the incidence of epidural abscess and spinal meningitis is generally so low that many of the available recommendations are based on evidence from other domains of infection control (e.g., surgical wound site and central venous catheter–related infection).^45–47 Nonetheless, neurologic com­promise resulting from neuraxial infection can be a devastating complication.

Infection of the epidural space tends to result in the formation of an abscess, most commonly formed by Staphylococcus aureus found in the epidermis of either the patient or the anesthesia provider. In contrast, meningitis associated with neuraxial procedures is most commonly caused by Streptococcus viridans. Viridans species of streptococcus may reside in the oronasopharyngeal tract of providers or patients or in the vagina. Potential routes of infection include the (1) epidural catheter track, (2) bloodstream, (3) equipment, and (4) injectate. A more in-depth discussion of neuraxial infection is found in Chapter 32.

Guidelines describing aseptic technique for regional and neuraxial anesthetic procedures have been published by professional anesthesiology organizations, including the ASA, the Association of Anaesthetists of Great Britain and Ireland, and the American Society of Regional Anesthesia and Pain Medicine.^47–49 The following recommendations deserve emphasis:

The first equipment decision involves determining whether to perform a single-shot or continuous technique. Continuous spinal anesthesia is not a new technique; indeed, some physicians performed continuous spinal anesthesia more than 50 years ago. Currently, a large-bore epidural needle and catheter must be used for continuous spinal anesthesia, because the U.S. Food and Drug Administration rescinded approval for the use of small-bore microcatheters in 1992.⁵⁴ Therefore, the risk for post–dural puncture headache is significant. This technique is useful after unintentional dural puncture with an epidural needle. In the morbidly obese patient, it may be easier to manipulate and advance a rigid epidural needle than a more flexible spinal needle; thus the technique is useful for establishing continuous analgesia or anesthesia in this patient population, particularly when the need for anesthesia is urgent. However, for most obstetric patients, a single-shot technique is preferred for spinal anesthesia.

The primary equipment choice for spinal anesthesia concerns the type and size of the spinal needle. Cutting-bevel needles (e.g., Quincke) are rarely used in contemporary obstetric anesthesia practice because of the unacceptably high incidence of post–dural puncture headache associated with their use.⁵⁵ Instead, non-cutting needles (e.g., Whitacre, Sprotte, Gertie Marx) are used almost exclusively (Figure 12-6). Some anesthesia providers refer to the non-cutting needles as “pencil-point needles.” It is now believed that the pencil-point needles cause more trauma to the dura, which then results in a more intense inflammatory response than occurs with cutting-bevel needles. Presumably, the inflammation results in more rapid closure of the dural defect.⁵⁶

With a small-gauge needle (i.e., 24-gauge or smaller), use of an introducer needle is preferable. The introducer needle engages the interspinous ligament and more accurately guides the trajectory of the smaller spinal needle than is possible with use of a small-gauge spinal needle alone. The introducer needle also aids with skin puncture; it is often difficult to puncture the skin with non-cutting needles.

Either the midline or the paramedian approach can be used to enter the subarachnoid space. The midline approach requires the patient to reduce her lumbar lordosis to allow access to the subarachnoid space between adjacent spinous processes (usually L3 to L4, sometimes L4 to L5 or L2 to L3). The interspinous space may be identified with one (usually the thumb or index finger) or two fingers (usually the index and middle fingers) of the anesthesia provider’s nondominant hand. The single finger “slides” along the skin in the midline from cephalad to caudad until it “settles” into an interspinous space. The two fingers identify the interspinous space by palpating the caudad border of the more cephalad spine. The fingers identify the midline by rolling in a medial-to-lateral direction (Figure 12-7).

Although with time and practice the tactile feedback produced by advancing a needle through tissues of variable resistance will become utterly familiar to the anesthesia provider, the novice may be unsure of the anatomic position of the needle tip, especially if unexpected resistance (i.e., contact with bone) or an unexpected and premature “pop sensation” is encountered during needle advancement. A stepwise problem-solving approach is reasonable. First, the anesthesia provider should reconfirm that (1) the patient has normal anatomy (i.e., not scoliotic), (2) she is acceptably positioned at an appropriate height for the anesthesia provider, and (3) the chosen point of needle insertion is the true midline plane. If these assertions are true, and the needle tip encounters bone, it is highly likely that the osseous structure is either the inferior or superior spinous process. One of two maneuvers may overcome this barrier. After slight withdrawal of the needle, simple angulation in a cephalad or caudad direction may redirect the needle trajectory sufficiently to achieve access to the central neuraxial canal. One must appreciate the “toughness” of the interspinous ligament. Even a 17-gauge epidural needle can be bent if the angle is changed without some prior retraction of the needle. Furthermore, if a spinal needle/introducer complex is used, care must be exercised that angulation of the spinal needle does not occur without first withdrawing it into the lumen of the introducer. Thereafter, the entire spinal needle/introducer complex is angulated before the spinal needle is re-advanced. Angulation of the spinal needle without first withdrawing it into the introducer creates a fulcrum at the junction of the introducer tip where the spinal needle emerges and can potentially damage or even shear the delicate spinal needle (Figure 12-8).⁵⁷

Alternatively, the needle may be withdrawn fully into subcutaneous tissue and either raised or lowered (cephalad or caudad) while still maintaining an angulation that is parallel to the original trajectory. Which approach is more effective may depend on the reason for the initial bone contact. If the patient has very narrow interspaces, then careful raising or lowering of the needle while maintaining a trajectory parallel to the floor may be appropriate. However, if the patient is overly flexed forward, it is possible that her lumbar spinous processes are projecting in a slightly upward angulation (relative to the perpendicular transverse plane). This may require that the needle be re-angulated accordingly. If bone is still encountered despite all these considerations, it is likely that the needle tip is in fact not in the midline plane and is contacting the vertebral lamina. This may occur if the initial skin puncture is not in the midline, the needle tip deviates from the midline as it is advanced, or the patient’s spine is rotated (either from poor positioning or scoliosis). Clues that the needle tip is not midline include (1) the patient complaining of lateralizing pain, (2) lack of CSF flow despite appropriate needle depth, and (3) the perception of “soft” or “mushy” tissue during needle advancement (paraspinous tissue) rather than the more “rigid” ligamentous tissue, or even a false “pop” as the needle tip exits the interspinous ligament laterally into paraspinous tissue. Much like the progressive modifications described earlier for correct alignment in the superoinferior plane, so too can these approaches be employed for redirecting in the lateral plane (Figure 12-9). The novice is advised to make systematic changes in a stepwise fashion, rather than indiscriminately changing needle direction without first considering the anatomic problem.

For most patients, the midline approach is faster and less painful than the paramedian approach. The midline approach is also easier to teach than the paramedian approach, because it requires mental projection of the anatomy in only two planes, whereas the paramedian approach requires appreciation of a third plane and estimation of the depth of the subarachnoid space from the skin (Figure 12-10).

A common error that is made with the paramedian approach is the insertion of the needle too far off the midline; the vertebral lamina then becomes a barrier to needle insertion. With the paramedian approach, the palpating fingers should again identify the caudad edge of the more cephalad spinous process. A skin wheal is raised 1 cm lateral and 1 cm caudad to this point; a longer needle is then used to infiltrate the deeper tissues in a cephalomedial plane. This step contrasts to the midline approach, in which the local anesthetic is not injected beyond the subcutaneous tissue. The spinal introducer is then inserted 10 to 15 degrees off the sagittal plane in a cephalomedial direction, and the spinal needle is advanced through the introducer needle toward the subarachnoid space. Another common error is to use an excessive cephalad angle with initial needle insertion. When the needle is inserted correctly and contacts bone, it is redirected slightly cephalad. If bone is again encountered, but at a deeper level, the slight stepwise increase in cephalad angulation is continued, and the needle is “walked” up and off the lamina. As with the midline approach, the characteristic feel of the ligamentum flavum and dura can be appreciated. The aim of the paramedian approach is to puncture the dura in the midline, even though the needle is inserted off the midline. Use of the paramedian approach requires insertion of a greater length of needle. Once CSF is obtained, the block is performed as it is with the midline approach.

Special equipment for epidural analgesia or anesthesia includes an epidural needle, an epidural catheter (for a continuous technique), and a loss-of-resistance syringe (for the loss-of-resistance technique to identify the epidural space). Single-shot epidural anesthesia is rarely used in obstetric practice, because the major advantage of epidural over spinal anesthesia is the ability to provide continuous anesthesia or analgesia without puncturing the dura with a large needle. An epidural needle with a lateral opening (e.g., Hustead, Tuohy) is most commonly used because it allows a catheter to be threaded through its orifice (Figure 12-11).

Two methods are used to identify the epidural space during needle advancement: (1) hanging drop method and (2) loss-of-resistance method. The majority of anesthesia providers use the loss-of-resistance method (Figure 12-12).

The traditional loss-of-resistance syringe is a finely ground glass syringe with a Luer-Lok connector. Plastic syringes are now available, and the choice is generally a matter of the anesthesia provider’s preference. The syringe is filled with 2 to 4 mL of saline, air, or saline with a small (0.25 to 0.5 mL) air bubble. There is some controversy regarding the use of air versus saline for detecting the point of loss of resistance.⁵⁸ Saline causes some syringe plungers to stick and may be confused with CSF during initiation of CSE anesthesia. Conversely, injection of air into the epidural space may contribute to patchy anesthesia,⁵⁹ and unintentional pneumocephalus may increase the risk for post–dural puncture headache.⁶⁰

Prior investigations have suggested superiority of loss-of-resistance with saline over air with respect to block success. This conclusion is further supported by the results of a meta-analysis that pooled six controlled trials (n = 1037) in which laboring women were randomized to receive loss-of-resistance to air versus liquid (saline or local anesthetic solution).⁶¹ This analysis suggested an increased risk for unblocked segments when loss-of-resistance to air was used, presumably owing to air bubbles leaking through the intervertebral foramina and residing adjacent to nerve roots.⁶²

Stay updated, free articles. Join our Telegram channel

Join