Spinal, Epidural, and Caudal Anesthesia

Neuraxial Anatomy

Anatomic Changes of Pregnancy

The normal anatomic changes of pregnancy affect the use of neuraxial anesthesia techniques. Uterine enlargement and vena caval compression result in engorgement of the epidural veins. Unintentional intravascular epidural catheter cannulation and injection of local anesthetic are more common in pregnant patients than in nonpregnant patients. In addition, the vertebral foraminal veins, which are contiguous with the epidural veins, are enlarged and obstruct one of the pathways for anesthetic egress from the epidural space during administration of epidural anesthesia. The enlarged epidural veins also may displace CSF from the thoracolumbar region of the subarachnoid space, as does the greater intra-abdominal pressure of pregnancy; this displacement partly explains the lowered dose requirement for spinal anesthesia in pregnant women.⁹ Subarachnoid dose requirements are also affected by the lower specific gravity of CSF in pregnant patients than in nonpregnant patients.¹⁰

The hormonal changes of pregnancy affect the perivertebral ligamentous structures, including the ligamentum flavum. The ligamentum flavum may feel less dense and “softer” in pregnant women than in nonpregnant patients; thus, sensing the passage of the epidural needle through the ligamentum flavum may be more difficult. It may also be more difficult for a pregnant woman to achieve flexion of the lumbar spine. Progressive accentuation of lumbar lordosis alters the relationship of surface anatomy to the vertebral column. At least three changes may occur. First, a pregnant woman's pelvis rotates on the long axis of the spinal column; thus, the line joining the iliac crests (Tuffier's line) assumes a more cephalad relationship to the vertebral column (e.g., this imaginary line might cross the vertebral column at the L3 to L4 interspace rather than the L4 to L5 interspace). Second, there is less space between adjacent lumbar spinous processes during pregnancy. It may be more difficult to use the midline approach to identify the epidural or subarachnoid space in pregnant women. (Thus the often-heard comment, “She has a narrow interspace.”) Third, magnetic resonance imaging has shown that the apex of the lumbar lordosis is shifted caudad during pregnancy, and the typical thoracic kyphosis in women is reduced during pregnancy.¹¹ These changes may influence the spread of subarachnoid anesthetic solutions in supine patients, leading to a higher sensory level in the pregnant patient (Figure 12-4).¹² Finally, labor pain makes it more difficult for some women to assume and maintain an ideal position while the anesthesia provider performs neuraxial anesthesia.

Obstetric Pain Pathways

Pain during the first stage of labor results primarily from changes in the lower uterine segment and cervix. Pain is transmitted by visceral afferent nerve fibers that accompany the sympathetic nerves and enter the spinal cord at the T10 to L1 segments. During the late first stage and second stage of labor, pain results from distention of the pelvic floor, vagina, and perineum. Pelvic pain is transmitted by somatic nerve fibers, which enter the spinal cord at the S2 to S4 segments (Figure 12-5).

During cesarean delivery, additional nociceptive pathways are involved in the transmission of pain, and a T4 level of anesthesia is required to provide adequate anesthesia. Most cesarean deliveries are performed with a horizontal (e.g., Pfannenstiel) skin incision, which involves the infraumbilical T11 to T12 dermatomes. During surgery, stretching of the skin may involve dermatomes two to four levels higher. Intraperitoneal manipulation and dissection involve poorly localized visceral pain pathways. Visceral pain may be transmitted by pathways as high as the celiac plexus. Additional somatic pain impulses may occur as a result of diaphragmatic stimulation because the intercostal nerves innervate a portion of the peripheral diaphragm.

Physiology of Neural Blockade

Hormonal changes, anatomic changes, and decreases in CSF specific gravity likely are responsible for the lower local anesthetic dose requirements for spinal anesthesia in pregnant women.^10,13 Local anesthetics produce conduction blockade primarily by blocking sodium channels in nerve membranes, thereby preventing the propagation of neural impulses. Differential blockade is manifested as differences in the extent of cephalad blockade of temperature discrimination and vasomotor tone, sensory loss to pinprick, sensory loss to touch, and motor function. Temperature discrimination and vasomotor tone are blocked to the greatest extent (i.e., most cephalad level) and motor function to the least extent. During spinal anesthesia, local anesthetics act directly on neural tissue in the subarachnoid space. Regression of anesthesia can be explained by the simple vascular uptake of local anesthetic from the subarachnoid space and spinal cord.¹⁴ Epidural anesthesia has a much smaller zone of differential motor–sensory–sympathetic blockade; this difference suggests that the mechanism of epidural anesthesia must involve more than simple diffusion across the dura. For many years, nerve fiber size was presumed to be the primary determinant of susceptibility to local anesthetic blockade (i.e., smaller fibers are blocked more readily than larger fibers). However, later studies have shown that the length of nerve fiber exposed to local anesthetic is as important as the size of the nerve fiber. Fink¹⁵ hypothesized that the length of nerve fiber exposed to local anesthetic affects the extent of the differential zone of motor and sensory blockade. With spinal anesthesia, the local anesthetic concentration required to block sufficient sodium channels to affect motor, sensory, and sympathetic function is less than that needed for the better-protected nerves found in the epidural space; thus, a wider band of differential blockade occurs during spinal anesthesia than during epidural anesthesia.

The understanding of the mechanisms of spinal and epidural anesthesia likely remains oversimplified. Nonetheless, it seems clear that spinal anesthesia results primarily from the effects of local anesthetic on the spinal cord, whereas epidural anesthesia results from the effects of local anesthetic on nerve tissue within both the epidural and subarachnoid spaces.

Pre-procedural Considerations

Equipment and Placement of Needle/Catheter

Ultrasonographic Guidance

Traditionally, surface anatomy visualization and palpation have been used to assess landmarks before initiation of the neuraxial procedure. This landmark-guided technique is effective for the vast majority of patients and, as a consequence, the use of ultrasonography has not played as prominent a role in neuraxial anesthesia as it has in peripheral nerve blockade. There exist, however, unique clinical circumstances in which immediate pre-procedural imaging of epidural anatomy can be highly beneficial. Such may be the case in patients with morbid obesity, derangements of spinal anatomy such as scoliosis or spinal stenosis, or a history of spinal instrumentation, and in patients in whom identification of specific vertebral levels might be warranted (e.g., known preexisting disc herniation or nerve root compression at a specific interspace). Unlike the use of ultrasonography for vascular access and peripheral nerve block techniques, ultrasonography for neuraxial techniques is not used in real time. Indeed, the real-time use of ultrasonography during administration of neuraxial anesthesia can be prohibitively cumbersome, is fraught with a tangible risk for violating aseptic technique, and raises unanswered questions about what untoward effects may result should ultrasound coupling medium (gel) breach neuraxial structures. Most often ultrasonography is a pre-procedural tool to aid the operator in the assessment of needle insertion site, needle angle, and estimated depth of the epidural space.

It remains to be seen whether neuraxial ultrasonography can significantly impact the clinical success of neuraxial procedures, whether that is measured in terms of time to completion of the spinal or epidural anesthesia procedure, or block success and patient satisfaction. Additionally, reductions in the incidence of complications of neuraxial anesthesia, such as unintentional dural puncture, have yet to be realized.

In the obstetric population, four randomized controlled trials have explored outcomes related to block success using the traditional landmark-based technique versus the ultrasound-aided technique.^85-88 Only one, however, divorced the unblinded ultrasonographer from those supervising the block and rendering assessments of success.⁸⁸ In this investigation,⁸⁸ 370 parturients receiving epidural labor analgesia were randomized to receive pre-procedural ultrasonographic imaging for determination of depth to epidural space or no imaging (control group). The estimated depth to the epidural space was conveyed to 15 first-year residents with little or no previous neuraxial anesthesia experience, who performed the blocks under the supervision of a staff anesthesiologist who was blinded to the group assignment. Obese patients were not excluded from this study, and there were no significant differences between groups in demographic factors. The authors found a significant reduction in the number of attempts needed to place the epidural catheter and in the need for catheter replacement in the ultrasonography group. There was no difference in the rate of unintentional dural puncture, although the study was underpowered for assessment of this outcome. The calculated number of epidural catheter placements with ultrasonography needed to avoid replacing one failed catheter (number needed to treat) was 26. Of note, the residents were not themselves randomized, nor were they tracked according to the number of cumulative procedures with and without ultrasonography. From the perspective of introducing trainees to the technique of neuraxial anesthesia procedures, ultrasonographic imaging may yet play a more substantial educational role.

How does pre-procedural ultrasonographic imaging affect neuraxial anesthesia success for patients anticipated to endure difficult needle placement? Chin et al.⁸⁹ identified 120 patients presenting for elective lower extremity orthopedic surgery who were predicted to endure difficult administration of spinal anesthesia. Their basis for qualifying these patients as “difficult” included a body mass index greater than 35 kg/m² with nonpalpable spinous processes, the presence of moderate to severe scoliosis, or a history of spinal surgery involving excision of two or more lumbar spinous processes. These subjects were randomized to receive pre-procedural ultrasonographic imaging of epidural anatomy or a conventional landmark-based technique. The total number of attempts required before successful dural puncture was twofold higher in the landmark control group than in the ultrasonography group. Although the average time spent completing the spinal anesthesia procedure was longer in the control group than in the ultrasonography group (mean = 7.3 minutes versus 5.0 minutes, respectively), if one accounted for the pre-procedural ultrasonographic examination, the total procedure time was shorter in the control group (7.9 minutes versus 12.2 minutes).

How accurately do pre-procedural ultrasonoanatomic measurements correlate to the actual distance from the skin to the epidural space? In general, although the correlation is high, ultrasonographic measurements of neuraxial anatomy may underestimate the true distance from skin to epidural space in the context of neuraxial procedures.^88,90-92 The factors that influence the small disparity between the two include (1) differences between the angulation of the imaging beam versus the angulation of the needle; (2) differences between the degree of exerted pressure and skin compression of the ultrasound probe versus the needle; (3) current lack of fidelity in the ability of ultrasonography to discriminate between ligamentum flavum, epidural space, and dura mater; and (4) operator dependency of typical onscreen measurement tools such as digital calipers. In a general obstetric population, the mean difference between the needle depth and ultrasonographic depth to the epidural space was 0.01 cm (95% confidence interval, −0.67 to 0.69 cm),⁹² whereas in a second study by the same investigators in which subjects were limited to obese parturients, the mean difference was 0.3 cm (95% confidence interval, −0.7 cm to 1.3 cm).⁹⁰

A low-frequency (2- to 5-Hz) curvilinear probe allows visualization of neuraxial structures beneath the skin. Low-frequency waves are preferable owing to the requisite depth of penetration. The curvilinear array allows enough ultrasonic scope to capture lateral structures such as the transverse processes. The ultrasound beam can be used to identify first the spinous processes if these are not palpable, then the interspinous spaces, and, finally, ligamentous structures. The ligamentum flavum and dura mater are dense tissues and will appear hyperechoic (white), like bone, whereas the less dense epidural and subarachnoid spaces will appear hypoechoic (black). A variety of imaging planes are now well described. Transverse and median sagittal as well as paramedian sagittal approaches have been documented. The transverse-interlaminar view is likely to be of most use when attempting neuraxial procedures using the midline approach. On the other hand, the paramedian sagittal view may be of use if a paramedian approach is anticipated (Figure 12-18). With the median sagittal approach, the spinous process will produce a shadow when the beam is placed directly over it, thus reducing the ability to appreciate any ligaments beyond it.

Intravascular Test Dose

The ideal method for excluding intravenous placement of the catheter is controversial. Intravascular placement of the epidural catheter may occur in as many as 7% to 8.5% of obstetric patients.⁹³ Failure to recognize intravenous placement of the epidural catheter and subsequent intravenous injection of a large dose of local anesthetic may lead to systemic local anesthetic toxicity, with CNS symptoms, seizures, cardiovascular collapse, and death.⁹⁴

The most common intravascular test dose contains epinephrine 15 µg. In normal volunteers, intravenous injection of epinephrine 15 µg (3 mL of a 1 : 200,000 solution) reliably causes tachycardia.^95,96 An increase in heart rate of 20 beats per minute (bpm) within 45 seconds was 100% sensitive and specific for intravascular injection in unpremedicated patients.⁹⁵ An increase in systolic blood pressure of between 15 and 25 mm Hg was also observed. Some anesthesiologists have expressed concerns about the use of an epinephrine-containing test dose in laboring women. Intravenous epinephrine may cause a transient decline in uterine blood flow as a result of alpha-adrenergic receptor–mediated constriction of the uterine arteries (Figure 12-19).^97-99

However, this decrease in uterine blood flow is transient and comparable to the decrease that occurs during a uterine contraction. Youngstrom et al.¹⁰⁰ noted that this intravenous dose of epinephrine did not worsen fetal condition in acidotic fetal lambs. In healthy parturients, any transient effect of epinephrine on uterine blood flow likely represents a less severe insult than systemic local anesthetic toxicity. An epinephrine-containing test dose, however, may not be appropriate in parturients with severe hypertension or uteroplacental insufficiency.

Some anesthesiologists argue that the epinephrine-containing test dose lacks specificity in laboring women. The maternal tachycardic response to intravenous injection of epinephrine cannot always be distinguished from other causes of tachycardia (e.g., pain during a uterine contraction) (Figure 12-20).^101,102

Cartwright et al.¹⁰¹ noted that 12% of laboring women had an increase in heart rate of at least 30 bpm after epidural injection of 3 mL of 0.5% bupivacaine without epinephrine. A study in laboring women compared the intravenous injection of epinephrine (10 to 15 µg) with that of saline; the sensitivity was 100%, the area under the receiver operator curve was 0.91 to 0.93, and the negative predictive value was 100%.¹⁰³ However, the positive predictive value was 55% to 73%. The results suggested that if a positive heart rate response to an epinephrine-containing test dose occurs in 20% of patients, 5% to 9% of epidural catheters would be identified incorrectly as intravascular and removed unnecessarily. Colonna-Romano and Nagaraj¹⁰⁴ concluded that the intravenous injection of an epinephrine-containing test dose results in “a sudden and fast acceleration in maternal heart rate within one minute.” Thus, careful assessment of the rate of increase in maternal heart rate may help distinguish a contraction-induced increase in heart rate from the effect of intravenously injected epinephrine, thereby improving the specificity of the epinephrine test. It is unclear whether such an assessment is clinically practical or will actually reduce the incidence of false-positive results.¹⁰⁵

Additionally, some anesthesiologists argue that the epinephrine-containing test dose lacks sensitivity (the ability to elicit a predictable increase in heart rate when the catheter is intravascular). An increase in maternal heart rate of 25 bpm occurring within 2 minutes of drug injection and lasting at least 15 seconds was observed in only 5 of 10 laboring women who received intravenous epinephrine 15 µg.¹⁰⁶ Detection of intravenous epinephrine injection was improved when the authors retrospectively defined a positive maternal tachycardic response as a 10-bpm increase above the maximum maternal heart rate observed in the 2-minute period preceding the epinephrine injection. Others have confirmed that these revised criteria improve the sensitivity of the epinephrine-containing test dose in laboring women.¹⁰³ These findings stress the importance of tracking the chronotropic variability that occurs in laboring patients during neuraxial anesthesia procedures to reduce the risk for misinterpretation of the test dose.

The usefulness of an epinephrine-containing test dose also improves if additional information is obtained. For example, investigators administered intravenous bupivacaine 12.5 mg with epinephrine 12.5 µg or saline to laboring women.¹⁰⁷ They correctly identified the test solution in 39 of 40 women when they assessed maternal heart rate, blood pressure, uterine contractions, the timing of the injection, the presence of analgesia, and subjective signs and symptoms of intravascular injection (e.g., palpitations, lightheadedness, dizziness). The tachycardic response to intravenous epinephrine is not a reliable indicator of intravascular injection in patients who have received a beta-adrenergic receptor antagonist.⁹⁵

Other means of identifying intravascular placement of an epidural catheter have been proposed and may be clinically useful in specific patients (Table 12-2). Intravenous administration of isoproterenol 5 µg consistently results in tachycardia in pregnant women.^108,109 Data from animals^99,110 and noninvasive measurements in parturients¹⁰⁹ suggest that isoproterenol is devoid of the adverse effects of epinephrine on uterine blood flow, and limited neurotoxicity evaluations have not revealed adverse effects.^110,111 However, isoproterenol has not been approved for epidural or intrathecal administration. Given the lack of adequate information regarding potential neurotoxicity, we do not recommend the use of isoproterenol as an epidural test dose.

Leighton et al.¹¹² have advocated the use of air as an objective marker of intravascular injection. Intravenous injection of 1 or 2 mL of air through a single-orifice catheter consistently produces changes in heart sounds as detected by the use of precordial Doppler ultrasonography.¹¹² (The external FHR monitor can be used for this purpose.) False-negative results may occur when small volumes of air are injected through a multi-orifice epidural catheter; thus, the air test is not a reliable test for intravascular injection when multi-orifice epidural catheters are used.¹¹³

Local anesthetic–induced symptoms of subclinical CNS toxicity have also been evaluated as a means of recognizing the unintentional intravenous injection of epidural medications. Colonna-Romano et al.¹¹⁴ administered intravenous saline, lidocaine 100 mg, or 2-chloroprocaine 100 mg to laboring women. Observers blinded as to which substance was administered recorded the presence of CNS symptoms (i.e., dizziness, tinnitus, funny taste) after intravenous injection of local anesthetic. Lidocaine 100 mg was a reliable marker of intravenous injection when the symptoms of tinnitus and funny taste were considered (sensitivity 100%; specificity 81%). 2-Chloroprocaine was less reliable (sensitivity 81% to 94%; specificity 69% to 81%). In a volunteer study, a dose of 1.5 mg/kg of 2-chloroprocaine was necessary to produce a probability of 90% that the subject would report symptoms of intravenous injection.¹¹⁵

Administration of fentanyl 100 µg has been described as a test for intravenous injection.¹¹⁶ Morris et al.¹¹⁷ evaluated the accuracy and reliability of the fentanyl test dose in a double-blind study in which either intravenous or epidural fentanyl 100 µg was administered to parturients, and the investigators sought evidence of the occurrence of sedation, dizziness, euphoria, and/or analgesia. Dizziness was the most reliable symptom of intravenous fentanyl injection, with a sensitivity of 92% and a specificity of 92%.

Some situations reduce the reliability of subjective symptoms as a signal of intravenous injection of a drug. Tests that rely on the self-reporting of subjective symptoms require clear communication with the patient and thus are less useful when the anesthesia provider and patient speak different languages. Patient exhaustion and/or prior opioid administration also may affect the reliability of the test.

Lastly, the epidural catheter design and speed of injection may affect the reliability of the epidural test dose. An epinephrine-containing test dose should be injected rapidly; otherwise rapid redistribution and metabolism of the drug decrease the actual dose administered to chronoreceptors. Multi-orifice epidural catheters have three potential sites of exit for injected fluid or air, and the orifices may lie within two different body compartments. If injected too slowly, air or fluid preferentially exits the proximal orifice. The speed of injection used in clinical practice typically exceeds that required to ensure that fluid will exit all three orifices. In contrast, air must be injected at a much greater speed to ensure that it exits all three orifices; this speed is not practical for clinical use. The distal orifice is both the most difficult to test and the one most likely to be positioned outside the epidural space.¹¹⁸

Intrathecal Test Dose

The intrathecal test dose should allow easy identification of subarachnoid (intrathecal) placement of the catheter without causing high or total spinal anesthesia and hemodynamic compromise. Bupivacaine (7.5 mg) and lidocaine (45 to 60 mg) are the local anesthetics most often used for an intrathecal test dose (see Table 12-2; Figure 12-21).^119,120

In a study of older, nonpregnant patients receiving continuous spinal anesthesia for surgery, Colonna-Romano et al.¹²⁰ used plain lidocaine 45 mg plus epinephrine 15 µg and assessed patient perception of lower extremity warmth and heaviness, sensory loss to pinprick, and ability to perform a straight-leg raise. Patients usually perceived warmth in their legs within 1 minute of the intrathecal injection; however, impaired straight-leg raise 4 minutes after an intrathecal test dose injection was the only test that had a sensitivity of 100% for intrathecal injection. The application of these data to pregnant patients is unclear.

Richardson et al.¹²¹ described the rapid onset (1 to 3 minutes) of high levels of spinal anesthesia with motor block and hypotension in five parturients who had received a test dose of plain lidocaine 45 mg plus epinephrine 15 µg. This solution is slightly hypobaric relative to CSF at body temperature; thus, the upright posture of these parturients during the injection may have contributed to the high levels of spinal anesthesia. The anesthesia provider must recognize the possible range of responses to the dose of local anesthetic used to assess the position of an epidural catheter and should perform a careful assessment of sensory, motor, and sympathetic function 3 to 5 minutes after administration of the test dose before concluding that the intrathecal test dose result is negative. Ropivacaine 15 mg is not a useful intrathecal test dose because the slow onset of motor blockade precludes timely diagnosis of intrathecal injection.¹²² 2-Chloroprocaine is the only local anesthetic that can be used as a single, combined intravascular and intrathecal test dose; in most patients, 100 mg results in dense, but not total, spinal anesthesia after intrathecal injection and produces systemic signs of subclinical toxicity (tinnitus, funny taste) after intravascular injection.

The epidural injection of local anesthetic for the purpose of ruling out intrathecal or intravascular catheter placement augments epidural analgesia and should be considered in the calculation of the initial therapeutic dose of local anesthetic. Several groups of investigators demonstrated that the test dose enhanced the density of epidural blockade and adversely affected the ability to ambulate. A test dose containing lidocaine 45 mg/epinephrine 15 µg adversely affected the ability to ambulate when injected immediately before initiation of epidural analgesia with 0.125% bupivacaine (18.75 mg) with sufentanil (10 µg).¹²³ Similarly, the same test dose interfered with ambulation when administered immediately after the intrathecal injection of bupivacaine 2.5 mg and fentanyl 25 µg.¹²⁴ Finally, a lidocaine 60 mg/epinephrine 15 µg test dose adversely affected the ability to ambulate in women who received neostigmine 500 µg with sufentanil 10 µg for initiation of analgesia.¹²⁵

Techniques to Minimize Local Anesthetic Toxicity

No perfect test dose exists. Some anesthesia providers elect not to administer a test dose because it contributes to motor blockade.^123,124 In addition, aspiration of a multi-orifice epidural catheter for blood has 98% sensitivity for detection of an intravascular location.⁶⁷ Inadvertent intravascular injection of a solution containing a low concentration of local anesthetic is unlikely to result in serious morbidity. However, this conclusion depends heavily on the use of small doses of local anesthetic.¹²⁶ Laboring women may need large doses of local anesthetic for operative delivery. In some cases, large doses are administered quickly (without incremental injection) for emergency cesarean delivery. The anesthesia provider wants to determine as soon as possible that the epidural catheter is correctly positioned within the epidural space. Even if no morbidity results, it is inconvenient for both the patient and the anesthesia provider to have to repeat the procedure and replace the catheter once the drape has been removed and the patient has been repositioned. The epinephrine-containing test dose provides an objective marker of intravascular injection that has stood the test of time. Thus, we typically give a test dose that contains either bupivacaine 7.5 mg or lidocaine 45 to 60 mg with 15 µg of epinephrine. No matter whether a test dose is injected, drugs should be injected incrementally into the epidural space, because no test is 100% sensitive and catheters may migrate during use. Each incremental dose should be treated as a “test dose” (i.e., the dose should be small enough that it will not cause systemic toxicity if unintentionally injected intravascularly or total spinal anesthesia if injected intrathecally). Pregnant women are very difficult to resuscitate from local anesthetic cardiovascular toxicity.⁹⁴

Steps to minimize the possibility of local anesthetic toxicity are summarized in Box 12-3. They include observation for passive return of CSF or blood (by lowering the proximal end of the epidural catheter below the insertion site), administration of the test dose between contractions, aspiration before each dose, incremental dose administration, maintaining verbal contact with the patient, and assessment for an appropriate level and density of sensory and motor blockade (which indicates correct placement of the catheter in the epidural space).

Anesthesia providers may give spinal anesthesia for cerclage, nonobstetric surgery during pregnancy, instrumental vaginal delivery, cesarean delivery, removal of a retained placenta, or postpartum tubal ligation. Spinal analgesia may be used for labor analgesia. Cesarean delivery represents the most common indication for spinal anesthesia in pregnant women. Most anesthesia providers administer a hyperbaric solution of local anesthetic for spinal anesthesia in obstetric patients. Use of a hyperbaric solution as compared with an isobaric solution results in a faster onset of block and a higher maximum sensory level with a shorter duration of blockade.¹²⁷ The urgency and anticipated duration of surgery dictate the choice of local anesthetic agent. The most common choice in the United States is bupivacaine. Other agents include lidocaine and tetracaine. Ropivacaine and levobupivacaine may be used but are not approved for spinal administration in the United States, and levobupivacaine is not available in the United States. Lidocaine provides a short to intermediate duration of action. Bupivacaine, tetracaine, levobupivacaine, and ropivacaine provide intermediate to long durations of action.

Anesthesia providers often add an opioid to the local anesthetic to improve the quality of anesthesia, particularly with regard to visceral stimulation, and to provide postoperative analgesia.^128,129 The addition of an opioid to the local anesthetic decreases the incidence of intraoperative nausea and vomiting. The short-acting, lipid-soluble opioids (i.e., fentanyl, sufentanil) contribute to intraoperative anesthesia, and morphine is often administered for postoperative analgesia. Epinephrine may be added to prolong block duration and perhaps improve block density. It was hoped that other adjuncts (e.g., clonidine, neostigmine) might allow for the administration of a lower dose of local anesthetic and thereby minimize sympatholytic side effects and hasten recovery. Side effects from these other adjuncts, however, have precluded their wide use in obstetric anesthesia practice (see Chapters 23, 26, and 28).

Epidural Anesthesia

Local anesthetic agents available for epidural administration in obstetric patients include 2-chloroprocaine, lidocaine, mepivacaine, bupivacaine, ropivacaine, and etidocaine. Mepivacaine and etidocaine are used infrequently in obstetric anesthesia practice.

Bupivacaine remains the most popular local anesthetic for analgesia during labor and vaginal delivery because of its differential sensory blockade, long duration of action, low frequency of tachyphylaxis, and low cost. Anesthesia providers infrequently administer bupivacaine for cesarean delivery because of the risk for cardiac toxicity and maternal mortality after unintentional intravascular injection of the drug.

Ropivacaine has gained popularity as an agent for epidural analgesia and anesthesia because it may result in less cardiac toxicity and greater differential sensory blockade than bupivacaine.¹³⁰

Levobupivacaine also has a more favorable safety profile than bupivacaine. Clinical trials have shown that ropivacaine and levobupivacaine have potency¹³¹ and analgesic qualities similar to those of bupivacaine,^132,133 with the probable exception of less motor nerve block.¹³⁴

Bupivacaine, ropivacaine, and levobupivacaine all have longer durations of action than lidocaine, and they may be preferred over shorter-acting agents when a longer duration of anesthesia or analgesia is desirable. They are more commonly used for maintenance of epidural labor analgesia, whereas the shorter-acting agents are used for epidural surgical anesthesia. Despite some variation among reports, published clinical studies suggest no more than slight differences in onset and potency, and no differences in quality or duration of neural blockade, among the three drugs. However, bupivacaine is more cardiotoxic than the other agents in vitro and probably after unintentional intravascular administration.¹³⁵ It would seem prudent to use ropivacaine or levobupivacaine rather than bupivacaine when a bolus dose of a concentrated solution is being given. When administered as a low concentration infusion, improved safety has not been demonstrated with ropivacaine and levobupivacaine compared with bupivacaine.

The most popular choice of local anesthetic for epidural anesthesia for cesarean delivery is 2% lidocaine with epinephrine. The addition of epinephrine (5 µg/mL) causes a modest prolongation of the block. The major advantage of epinephrine is that it improves the quality of epidural lidocaine anesthesia. Lam et al.¹³⁶ have shown that epidural labor analgesia can be extended to surgical anesthesia for cesarean delivery in 5.2 ± 1.5 minutes (mean ± SD) with the addition of bicarbonate and fentanyl to 2% lidocaine with epinephrine.

Many anesthesia providers reserve 2-chloroprocaine for cases in which rapid extension of epidural anesthesia for vaginal delivery or urgent cesarean delivery is necessary. The onset of surgical anesthesia was several minutes faster with 2-chloroprocaine compared with lidocaine with freshly mixed epinephrine and sodium bicarbonate in the setting of urgent caesarean delivery after epidural labor analgesia.¹³⁷ Therefore, when time is of the essence, 2-chloroprocaine is the drug of choice. Typically, in an emergency, a large volume of concentrated local anesthetic solution is injected quickly. An additional advantage of 2-chloroprocaine in this situation is that it is rapidly metabolized by plasma esterases. Therefore, the unintentional intravascular injection of a large volume of 2-chloroprocaine may be less likely to have serious adverse consequences. A potential disadvantage of 2-chloroprocaine is that it may interfere with the subsequent actions of opioids¹³⁸ and bupivacaine,¹³⁹ although this possibility is controversial.¹⁴⁰

As in spinal anesthesia, epidural opioids work synergistically with local anesthetics. Fentanyl 50 to 100 µg or sufentanil 5 to 10 µg is frequently added to an amide local anesthetic for both labor analgesia (allowing a lower dose of local anesthetic and less motor block) and cesarean delivery (resulting in a denser block with better blockade of visceral stimulation). Sodium bicarbonate may be added to lidocaine¹⁴¹ and 2-chloroprocaine¹⁴² (1 mEq/10 mL local anesthetic) to decrease latency.

Caudal Anesthesia

The drugs used for caudal epidural anesthesia are identical to those used for lumbar epidural block. However, a much larger volume (e.g., 25 to 35 mL) of local anesthetic solution must be administered to extend a caudal block for cesarean delivery or labor analgesia. Such large volumes entail a greater risk for systemic local anesthetic toxicity. Additionally, the caudal epidural space is highly vascular, which further predisposes to intravasation of drugs administered via this route.

Unintentional Dural Puncture

Unintentional dural puncture with an epidural needle occurs at a rate of approximately 1.5% in the obstetric population.¹⁴³ Approximately 52% of women will experience a post–dural puncture headache after puncture with an epidural needle (see Chapter 31). Techniques to minimize the incidence of unintentional dural puncture include (1) identification of the ligamentum flavum during epidural needle advancement; (2) understanding the likely depth of the epidural space in an individual patient; (3) advancement of the needle between contractions, when unexpected patient movement is less likely; (4) adequate control of the needle-syringe assembly during advancement of the needle; and (5) clearing the needle of clotted blood or bone plugs. Norris et al.¹⁴⁴ observed that post–dural puncture headache after unintentional dural puncture was less likely to result in headache if the epidural needle bevel faced lateral rather than cephalad. In contrast, Richardson and Wissler¹⁴⁵ found no difference between the two orientations. An in vitro study using cadaver dura found that the fluid leakage rate through dural tears was not dependent on the orientation of the dura relative to the needle bevel.¹⁴⁶ We prefer to insert the epidural needle with the bevel oriented in a cephalad direction so that there is no need to rotate the needle bevel within the epidural space. Cephalad bevel orientation also increases the likelihood of successful epidural anesthesia.¹⁴⁷

The management of unintentional dural puncture depends on the clinical setting. If advancement of an epidural needle results in dural trespass and the free flow of CSF is perceived in the barrel of the epidural syringe, the anesthesia provider should halt advancement of the epidural needle and either remove the epidural needle if the plan is to resite the epidural catheter or replace the stylet back into the lumen of the epidural needle to prevent further egress of CSF if the plan is to administer intrathecal medication or perform a continuous spinal technique. Reinjection of CSF contained in the syringe should not be entertained because there is a high likelihood that air will concomitantly be introduced into the subarachnoid space and cause pneumocephalus. The risks and benefits of administering an intrathecal dose of anesthetic should be considered. Although intrathecal administration of local anesthetic through the epidural needle will result in rapid analgesia for an uncomfortable patient, the rapid efflux of CSF may render the injectate ineffective. Additionally, spinal analgesia may mask paresthesias during subsequent attempts at neuraxial anesthesia. Furthermore, spinal anesthesia may cause profound hypotension. If the patient is sitting, and likely to remain so for the next few minutes, the anesthesia provider may struggle to manage two problems instead of one (i.e., a challenging neuraxial procedure and maternal hypotension). However, rapid provision of analgesia may enable the patient to better assume an optimal position without moving. One option is to administer intrathecal opioid, thus providing analgesia without the risk for hypotension.

The anesthesia provider has two options after unintentional dural puncture: site the epidural catheter within the subarachnoid space and use a continuous spinal anesthetic technique or site the epidural catheter in a different interspace (i.e., starting afresh). Evidence is conflicting as to whether the insertion of an epidural catheter through the dural puncture site decreases the incidence of post–dural puncture headache.¹⁴⁸ Continuous spinal anesthesia is an attractive option if identification of the epidural space has been difficult or if the anticipated duration of epidural anesthesia or analgesia is relatively short (e.g., cesarean delivery, or vaginal delivery in parous women). The major disadvantage of an intrathecal catheter is the risk that it may be mistaken for an epidural catheter. Given that the local anesthetic dose required for epidural anesthesia is many times greater than that required for spinal anesthesia, unintentional administration of an epidural dose into the subarachnoid space will lead to total spinal anesthesia. Therefore, on a busy labor and delivery unit where multiple providers are giving anesthesia care, it may be safer to use an epidural catheter rather than an intrathecal catheter in women in whom prolonged analgesia is anticipated.

Thus, the anesthesia provider may elect to initiate epidural anesthesia in another lumbar epidural interspace. However, even if the attempt results in a correctly placed catheter, the provider must be wary of an unexpectedly high level of anesthesia after the epidural administration of usual doses of local anesthetic.^149,150 Leach and Smith¹⁵⁰ reported a patient who had an extensive block after unintentional dural puncture and subsequent epidural injection of bupivacaine. They presented radiologic evidence of the spread of local anesthetic from the epidural space to the subarachnoid space. The extent to which a dural tear affects the movement of substances from the epidural space to the subarachnoid space depends on the size of the hole, the lipophilicity of the drug (highly lipophilic drugs cross quickly regardless of the presence of a hole, whereas water-soluble drugs cross more quickly in the presence of a hole),¹⁵¹ and whether the drug is administered into the epidural space as a bolus or an infusion. Rapid bolus administration of medications through an epidural catheter in a patient with a known dural puncture with a large-bore needle significantly increases the likelihood of high spinal anesthesia.

Unfortunately, there is no reliable method to decrease the risk for post–dural puncture headache once dural puncture occurs. Obese patients appear to be at lower risk than lean patients for the development of headache.¹⁵² A prophylactic blood patch (injection of autologous blood before removal of the epidural catheter and before onset of a headache) does not reduce the risk for post–dural puncture headache.^148,153

Occasionally, unintentional dural puncture is not recognized until the epidural catheter is threaded and CSF spontaneously appears at the proximal end of the catheter, or the catheter aspiration or intrathecal test dose is positive. Finally, an epidural catheter that has been correctly sited in the epidural space may migrate into the subarachnoid space. The most significant clinical threat in this scenario is the continued use of a large volume infusion of local anesthetic drug intended for epidural administration. Thus, during prolonged epidural labor analgesia, the patient should be monitored for evidence of high or dense neuraxial anesthesia.

Unintentional Intravascular or Subarachnoid Injection

The unintentional injection of large doses of local anesthetics into the subarachnoid space can lead to catastrophe. The rapid onset of high or total spinal anesthesia results in profound hypotension, loss of consciousness, and apnea secondary to hypoperfusion of the brain stem. Prompt treatment necessitates assisted ventilation, volume resuscitation, and pharmacologic support of blood pressure. Administration of chronotropic agents such as epinephrine may also be necessary if blockade of cardiac sympathetic drive results in bradycardia. The patient is at high risk for awareness in this setting, and the judicious use of an amnestic agent such as midazolam should be considered once cardiorespiratory stability has been restored.

The incidence of intravascular catheter placement varies according to catheter type,^33,65 patient population,¹⁵⁴ and proper placement of the epidural needle tip in the midline. Pregnant women are at higher risk for unintentional intravenous cannulation due to the engorgement of epidural veins. Intravascular injection of a local anesthetic may initially result in altered sensorium, tinnitus, and perioral numbness. Higher blood concentrations may result in seizures, and even higher concentrations may cause dysrhythmias and cardiovascular collapse (see Chapter 13). Management of intravascular toxicity includes prompt institution of advanced cardiac life support (ACLS), gamma-aminobutyric acid (GABA)–potentiating agents such as benzodiazepines to mitigate seizure activity, and the use of intralipid emulsion to detoxify the patient. Several modifications have been made to the conventional ACLS protocol for the specific treatment of local anesthetic-induced cardiac arrest.¹⁵⁵ These include using small, judicious doses of intravenous epinephrine for circulatory support (10 to 100 µg boluses in adults), avoidance of vasopressin, and the use of amiodarone in place of local anesthetics for treatment of ventricular arrhythmias.

Inadequate Anesthesia

Pain during anesthesia represents a higher proportion of obstetric malpractice claims than of nonobstetric claims.¹⁵⁶ During labor, inadequate epidural analgesia may result from the inadequate extent of sensory blockade, nonuniform blockade, or inadequate density of blockade. When called to evaluate breakthrough pain, the anesthesia provider should first evaluate the extent of bilateral sensory blockade in both the cephalad and caudad directions. Particularly if labor is progressing quickly, the extent of sacral blockade may not be adequate. In this case, epidural injection of a large volume of local anesthetic may improve sacral blockade. In contrast, if the extent of sensory blockade is adequate but the patient is still experiencing pain, the density of blockade may be insufficient. In this case, the provider should reestablish and maintain analgesia using a more concentrated solution of local anesthetic.

Why do some obstetric epidural anesthetics fail over time? Collier¹⁵⁷ administered epidural radiocontrast dye in 25 parturients reporting unsatisfactory analgesia. The two major causes of inadequate block in this small study were transforaminal migration of the catheter tip and an obstructive barrier in the epidural space. Total block failure usually results from failure to identify the epidural space correctly or from malposition of the catheter tip outside the epidural space (e.g., in a neuroforamen). A unilateral block may occur despite the use of good technique. Unilateral blocks can often be prevented by limiting the length of catheter within the epidural space to 3 cm or less. The problem with limited insertion of the catheter is that, in some patients, the catheter tends to migrate outward over time. (Patients undergoing surgery remain still; by contrast, laboring women change position frequently.) Obese women seem to be at higher risk for outward migration of the catheter tip. Prospective studies suggest that 4 to 6 cm is the optimal depth of epidural catheter insertion in laboring women.^71,72,158

Whether catheter withdrawal in the setting of breakthrough pain is beneficial is not clear. Beilin et al.¹⁵⁹ compared catheter withdrawal followed by injection of local anesthetic with injection of local anesthetic without catheter withdrawal for the treatment of breakthrough pain. The ability to rescue analgesia was not different between the groups. Gielen et al.¹⁶⁰ performed a radiologic study in which they observed no consistent relationship between catheter position and the asymmetric onset of sympathetic blockade. Unilateral or patchy sensory blockade likely results from the nonuniform distribution of local anesthetic solution in the epidural space.¹⁶¹ Injection of a large volume of dilute local anesthetic solution into the epidural space usually corrects this problem, regardless of the location of the tip of the epidural catheter (provided it is actually in the epidural space). If analgesia cannot be rescued with a second injection, the catheter should be removed and replaced at another interspace.

The management of inadequate anesthesia is more problematic during cesarean delivery. Failure of spinal anesthesia may result from the maldistribution of local anesthetic within the subarachnoid space.^162,163 If inadequate spinal anesthesia is noted before incision, the anesthesia provider may augment the block with additional local anesthetic by either performing a second spinal anesthetic procedure or placing an epidural catheter. However, care must be taken if performing a second spinal anesthetic procedure. In the ASA Closed-Claims Project database, Drasner and Rigler¹⁶² identified three cases of cauda equina syndrome complicating spinal anesthesia. In two cases, “failed spinal” anesthesia had occurred, followed by a repeat injection of local anesthetic. The researchers recommended that anesthesia providers determine the presence of anesthesia in the sacral dermatomes before giving additional local anesthetic into the subarachnoid space. Additionally, they stated that if CSF was aspirated during the original procedure, it should be assumed that local anesthetic was delivered into the subarachnoid space, and the total dose of local anesthetic be limited to the maximum dose a clinician would consider reasonable to administer in a single injection.¹⁶² If partial blockade is present (even if it is limited to the sacral dermatomes), the second dose should be reduced accordingly. It may also be advisable to perform the second procedure at a different interspace or make other changes to the original procedure (e.g., alter the patient's position, use a local anesthetic with different baricity, or straighten the lumbosacral curvature).

If the patient complains of pain after incision, the anesthesia provider must decide between the administration of inhalation or intravenous analgesia and the administration of general anesthesia. Supplemental analgesia may be provided by giving 60% nitrous oxide in oxygen, small incremental boluses of ketamine (0.1 to 0.25 mg/kg), or small boluses of intravenous opioid. Supplemental infiltration of the wound with local anesthetic is sometimes helpful, especially when spinal anesthesia regresses near the end of an unexpectedly long operation. The anesthesia provider must ensure that the patient remains sufficiently alert to protect her airway. In most cases, severe pain unrelieved by modest doses of analgesic drug requires rapid-sequence induction of general anesthesia followed by endotracheal intubation and general anesthesia.

In some cases, inadequate epidural anesthesia results from failure to give a sufficient dose of local anesthetic or failure to wait a sufficient time after its administration. For example, after the epidural administration of 0.5% bupivacaine, approximately 20 minutes must pass to achieve an adequate level of anesthesia, and additional local anesthetic may be needed to achieve an adequate density of blockade. In urgent cases or in cases with a “missed” segment, local infiltration with a local anesthetic often results in satisfactory anesthesia. Sometimes it is difficult to separate the beneficial effect of the local infiltration from the beneficial effect of waiting for the obstetrician to obtain, prepare, and inject the local anesthetic solution. Finally, the anesthesia provider should exercise caution when initiating spinal anesthesia after failure of epidural anesthesia because of a greater incidence of high spinal anesthesia in this setting.¹⁶⁴ Presumably, the large volume of local anesthetic within, or near, the epidural space results in decreased lumbar CSF volume, which predisposes to high spinal anesthesia. It may be advisable to reduce the dose of intrathecal local anesthetic, particularly in the presence of partial epidural blockade.

Equipment Problems

The frequency of major equipment malfunction is very low during the administration of neuraxial anesthesia. Most anesthesia providers in the United States use disposable needles, and the plastic needle hubs are attached to the needles' shafts with epoxy. Rarely, a needle breaks at the hub-shaft junction.¹⁶⁵ If a needle should break, the portion of the needle that remains in the patient should be removed, because it may migrate and cause injury.

An epidural or spinal catheter may shear and break off if the catheter is withdrawn through a needle; thus an epidural or spinal catheter should never be withdrawn in this manner. Rather, if the catheter must be withdrawn, the needle and catheter should be withdrawn as a unit. It is also possible to break a catheter during attempts at removing it, although this is rare. If resistance to catheter removal is encountered, the patient should assume a position that reduces lumbar lordosis, thereby lessening the kinking of the catheter between perivertebral structures. If position change is not successful, the catheter should be taped under tension to the patient's back and left undisturbed for several hours. The catheter usually works its way out and is then easy to remove. Once the catheter has been removed successfully, it should be examined to ensure that it has been removed completely. Complete removal of the catheter should be documented in the medical record.

Rarely, catheters do break on removal. We favor aggressive attempts to remove broken spinal catheters. However, it may be unnecessary to remove broken epidural catheters; rather, in these circumstances, the patient can be informed of the complication and observed over time. The incidence of catheter migration or other delayed sequelae appears to be low. Computed tomography may help identify the precise location of a broken catheter.¹⁶⁶

During use, an epidural catheter occasionally becomes disconnected from the catheter connector. Options include replacing the epidural catheter or reconnecting the connector to the catheter. Langevin et al.¹⁶⁷ used an in vitro model to investigate whether microbial contamination precludes reconnection. They found that an area of the interior of the catheter distal to the disconnection may remain sterile for up to 8 hours if the fluid column within the catheter remains static (i.e., if “fluid does not move within the catheter when it is raised above the level of the patient”).¹⁶⁷ Therefore, they concluded that it may be safe to decontaminate the exterior of the catheter, cut the catheter with a sterile instrument, and reconnect it to a new sterile connector. However, given the potential catastrophic consequences of neuraxial infection, we recommend replacing the catheter. Also, wire-embedded catheters cannot be cut.