Spinal Anesthesia

Spinal vs. Epidural

Spinal Epidural
Injection Location lumbar only anywhere
Duration of Block brief prolonged
Procedure Time brief longer
Quality of Block high not as good as spinal
Disadvantages increased risk of hypotension, dural puncture headache
Advantages ability to produce segmental block, greater control over analgesia, possibility of long term analgesia

Note that profound muscular blockade occurs with neuraxial anesthesia, obviating the need for NMBDs

Anatomical Points

Curvature is key in spinal anesthetics (no matter in epidurals). Note that in the lumbar area, the spinous processes are near-perpendicular to the VB, whereas in the thoracic area they point downwards. Also in the thoracic area, the interlaminar space is only a few millimeters. The sacral hiatus (unfused opening between S4 and S5) is missing in 8% of adults. C7 is the bony knob at the bottom of the neck. T7-8 is at the lower limits of the scapulae. Terminal point of 12th ribs is at L2. The line across the iliac crests crosses L4 VB. Posterior iliac spines are at S2 (caudal limit of dural sac in adults). The cord itself terminates at L1 in adults and at L3 in infants.

Vertebral Level Landmark
C7 Bony knob at base of neck
T7-8 Lower limits of scapulae
L2 Terminal point of 12th ribs
L4 Line across iliac crests
S2 Posterior iliac spines

Dura mater thins as nerves exit the intervertebral canal, facilitating penetration of local anesthetic. Spinal subarachnoid space is continuous with intracranial, thus excessive migration can lead to blockade of cranial nerves. The epidural space is not a closed space and communicates with the paravertebral spaces via the foraminae. The depth of the epidural space is maximal at L2 (terminal point of 12th ribs), where it is 6 mm in depth. It is 4-5 mm in the midthoracic area. There is considerable debate about whether or not the plica mediana dorsalis, which is purported to connect the dura mater to the ligamentum flavum, even exists, and if so, how relevant it is [Harrison Br J Anaesth 83: 229, 1999]. Artery of Adamkiewicz is highly variable but most commonly enters the canal at the left L1 foramen. The internal venous plexus, which drains the cord, is prominent in the lateral epidural space and ultimately empties into the azygous system.

Preoperative Preparation


Complications to discuss with the patient include 1) nerve damage 2) bleeding 3) infection 4) headache 5) failed block.


Indications for spinal anesthesia include lower abdominal, perineal, and LE surgery. Technically one could use it for upper abdominal surgery however because these procedures impact breathing so profoundly, general anesthesia is generally preferred. Indications for epidural anesthesia include abdominal or lower extremity surgery, however because of its segmental nature, it may be suboptimal for procedures involving the lower sacral distribution. Epidural anesthesia is also often used as a supplement to general anesthesia, as well as for labor pain.


Contraindications to neuraxial anesthesia include patient refusal, infection, bleeding diathesis, and ICP (?). Bacteremia is only a relative contraindication – risks and benefits must be weighed, but there is evidence from animal models that giving antibiotics prior to the procedure can reduce infection [Carp et. al. Anesthesiology 76: 739, 1992]. Other relative contraindications include cardiac disease (avoid acute decreases in SVR), and abnormal coagulation studies (?)

There is considerate controversy regarding the use of neuraxial anesthesia in patients with pre-existing neurologic disorders, such as multiple sclerosis. Baby Miller recommends avoiding it unless absolutely necessary.

Spinal Anesthesia

Monitor with capnography if possible. Generally, a preoperative opioid is helpful in relieving the pain associated with needle insertion, although local anesthetics are often adequate. Anatomically, the sitting position is preferable, however in heavily sedated patients can lead to vasovagal syncope.

Failure at the L4-5 interspace is 7% [Munhall RJ et al. Anesth Analg 67: 843, 1988] and decreases as one moves cephalad. Note that the cord reaches L3 in 2% of adults, and that using the iliac crests as a results in selection of an intervertebral space above L4 in 51% of cases [Broadbent et. al. 55: 1122, 2000], thus one should probably not attempt a spinal above the L3-4 interspace. [Reynolds Anaesthesia 56: 238, 2001]

Post-dural puncture headache is affected by needle selection – pencil point (Whitacre or Sprotte) needles have a lower incidence than beveled (Quincke) tips [Buettner et. al. Reg Anesth 18: 166, 1993], thus a 24 or 25 ga. pencil-tipped needle is most often used in younger people (i.e., more likely to develop PDPHA). Pencil-tipped needles do require more force than beveled needles, however they also provide more tactile information. When using a pencil-tipped needle with a side-port (i.e., Whitacre or Sprotte), direct the opening wherever you want the flow to go (note that for beveled needles the direction does not affect flow). Though controversial, some believe that if a bevel needle is used, orienting in a longitudinal direction reduces PDPHA. [Mihic DN. Reg Anaesth 9: 54, 1986]

The midline approach is easiest and passes through less sensitive structures. The paramedian approach is better suited to narrow interspaces or difficulty with flexion – typically start 1 cm from midline. The most common error with this approach is to underestimate the distance to the dura and cross the midline early. For VERY difficult cases, the Taylor approach can be used – start 1 cm medial and caudal to the PSIS, and advance cephalad at 55 degrees with a medial orientation based on the width of the sacrum – this approach is difficult but completely independent of patient flexion.

Level and duration are primarily determined by 1) baricity 2) contour of spinal canal 3) patient position in the first few minutes after injection. When lumbar lordosis is non-optimal, either place a pillow under the patient’s knees or place him/her in the lateral position. Isobaric solutions undergo less spread than hyperbaric solutions and actually behave as slightly hyperbaric because of their low temperature – these solutions are suited for perineal or lower extremity surgery. Hypobaric solutions (sterile water or 1/2 NS, the latter of which causes less osmotic stress to neural tissues) are rarely used, only in prone jackknife positions or with patients undergoing hip arthroplasty.

Spread is affected by the addition of vasoconstrictors (usually 0.1 – 0.2 mg of epinephrine i.e., 0.2 to 0.5 cc of 1:1000, or 2 – 5 mg phenylephrine). Epinephrine may have the additional benefit of some alpha-2 related analgesia. Tetracaine receives the most benefit from vasoconstriction.

Local Anesthetics

Tetracaine: major vasodilation (vasopressors have a profound affect)
Lidocaine: less vasodilation
Bupivacaine: decreases spinal and dural blood flow

Note that the addition of epinephrine to lidocaine has been associated with neurotoxicity in animal studies [Hashimoto et. al. Anesthesiology 94: 876, 2001] and was used in some of the cases of toxicity reported in the literature [Rigler ML et. al. Anesth Analg 72: 275, 1991; Drasner et. al. Anesthesiology 77: 582, 1992; Gerancher J. Anesthesiology 87: 687, 1997]. Furthermore, the addition of vasoconstrictors to tetracaine has been associated with increased transient neurologic symptoms [Freedman et. al. Anesthesiology 89: 663, 1998; Smith KN et al. Anesth Analg 98: 81, 2004]

Choice of Local Anesthetic – Short Duration Spinal Procedures

Chloroprocaine: initially linked to neurologic injuries in the 1980s, later determined that these were due to either the preservative or accidental injection of epidural doses. Recent studies suggest that low-dose chloroprocaine (40-60 mg) produce excellent short-duration spinal anesthesia. The use of vasoconstrictors with chloroprocaine has produced side effects, thus vasoconstriction is contraindicated. Fentanyl and clonidine have been shown to provide significant enhancement with no side effects. Will likely replace lidocaine as the drug of choice for short procedures.

Lidocaine: duration 60-90 minutes. Good sensory and motor block. Favorable recovery profile. Dose is now 60-75 mg, diluting the 5% stock solution with equal part saline (or CSF). Has been linked to TNS in up to 1/3 of patients receiving lidocaine for spinal anesthesia [Hampl K et al. Anesth Analg 81: 1148, 1995]. Outpatient status seems to increase the risk.

Choice of Local Anesthetic – Longer Duration Spinal Procedures

Bupivacaine and tetracaine are most common (ropivacaine has been used as well but does not seem to offer any advantage).

Bupivacaine: similar dose and duration as tetracaine (5-20 mg, 90-120 mins), slightly more intense sensory anesthesia (and less motor blockade) than tetracaine.

Tetracaine: similar dose and duration as bupivacaine (5-20 mg, 90-120 mins), slightly more motor blockade (and less sensory anesthesia) than bupivacaine. Duration is more variable than bupivacaine and more profoundly affected by vasoconstrictors.

Addition of Opiates

Opiates can be added (usually 25 ucg fentanyl) and affect the dorsal horn. Morphine (0.1 – 0.5 mg) can be used and provides 24 hours of relief, but unlike fentanyl, requires in-hospital monitoring for respiratory depression. Clonidine is sometimes added but is not as effective as the opiates [Eisenach et. al. Anesthesiology 85: 655, 1996]

Timing of Anesthesia

When giving a spinal anesthetic, the first 5-10 minutes are critical in terms of monitoring the cardiovascular response as well as the level. Temperature changes are the first to go, and a wetted alcohol swab can give an early indicator of the level (30-60 seconds), allowing the anesthesiologist to change the patient’s position if need be.

Sensory Level Type of Surgery
S2-S5 Hemorrhoidectomy
L2-L3 Foot surgery
L1-L3 Lower extremity
T10 (umbilicus) Hip, TURP, vaginal delivery
T6-T7 (xiphoid) Lower abdomen, appendectomy
T4 (nipple) Upper abdomen, C-section

Note that maldistribution is a common cause of failure, and that re-administration of a second, full dose may increase the risk of injury. [Drasner et. al. Anesthesiology 75: 713, 1991]

Physiology of Spinal Anesthesia

Spinal anesthesia blocks small, unmyelinated sympathetic fibers first, after which it blocks myelinated (sensory and motor) fibers. The sympathetic block can exceed motor/sensory by two dermatomes. Spinal anesthesia has little effect on ventilation but high spinals can affect abdominal/intercostal muscles and the ability to cough. Patients may complain of dyspnea because they can’t feel themselves breathing. Anything above T5 inhibits SNS to the GI tract. Some operations (hip, TURP) may bleed less during neuraxial blockade due to decrease systemic blood pressure. Some procedures (hip) may suffer less VTE due to increased blood flow to the lower extremities.

Side Effects of Spinal Anesthesia

Local anesthetics have been shown to produce permanent injury [Rigler et al. Anesth Analg 72: 275, 1991; Drasner et. al. Anesthesiology 75: 713, 1991]. Hypotension occurs in 1/3 of patients, initially due to decreased SVR but in severe cases due to decreased venous return and cardiac output (GREATLY enhanced by hypovolemia). Baby Miller recommends a modest head-down position (5-10 degrees) to increase venous return without altering the spread of anesthetic. Hydration is critical, although in excess can be detrimental. Ephedrine is the first line drug (phenylephrine may decrease cardiac output but is still commonly used by anesthesiologists, may have a role in an add-on drug when ephedrine causes increased HR). 10-15% of patients will experience bradycardia, the treatment of which is volume -> ephedrine -> atropine -> epinephrine as needed.

Post-dural puncture headaches are postural and can be accompanied by abnormalities on formal audiographic testing. Risk factors include age (peaks slightly after puberty, children and older people are rare), needle type (24-25G pencil point tips are ideal), and possibly gender (although the incidence of PDPHA in women may simply reflect the vulnerability of pregnant women [Lybecker H et al. Anesth Analg 70: 389, 1990]). Treat with bed rest, IVF, analgesia, caffeine, and possibly a blood patch (15-20 mL, injected at or below the site, as the blood will travel cephalad).

High spinals are often accompanied by hypotension, nausea, and agitation. “Total spinal anesthesia” is accompanied by LOC. Treat with ABCs (airway control and ventilation, IVF, sympathomimetics). Nausea which occurs after a spinal alerts the physician to the possibility of a high spinal and hypotension severe enough to cause a stroke, thus nausea is a critical warning sign, although it can also be caused by a predominance of residual parasympathetic activity.

Other potential side effects include urinary retention, backache, and hypoventilation secondary to thoracic or cervical spread.

Epidural Anesthesia

(also see article on Epidural_Analgesia)

Placement Technique

There is considerable controversy about placing these catheters after general anesthesia, and a retrospective review of this issue (no neurologic complications in 4298 patients undergoing lumbar epidural catheter placement while under general anesthesia for thoracic surgery at the Mayo Clinic) challenges the old assumption that the risks are greater [Horlocker TT et al. Anesth Analg 1547: 96, 2003], as does the practice of placing epidurals in asleep children [Krane et. al. Reg Anesth Pain Med 23: 433, 1998]. That said, most do not place these in unconscious patients for a variety of reasons. [Rosenquist et. al. Anesth Analg 96: 1545, 2003].

Patients usually receive some form of sedation prior to insertion. Tuohy (blunt-tip, gentle curve) needles are most commonly used and help guide the catheter direction. Multiorifice catheter enhance distribution but require greater depth. The depth of the epidural space is usually 4-6 cm, thus “finder needles,” which are 3.8 cm in length, are often (but not always) not deep enough.

For lumbar epidurals, the midline approach relies on understanding simpler anatomy, and also passes the needles through less sensitive structures than the paramedian approach (which can place the needle near the facet joints and spinal nerves).

For thoracic epidurals, the paramedial approach is more common (0.5 – 1.0 cm off midline). Finder needle is used to contact the lamina and inject local there. This is repeated with the epidural needle, which is then incrementally moved medial and cephalad.

The epidural space can be identified by either loss of resistance (passage through ligamentum flavum) or the hanging drop technique, although the latter is likely to be associated with a higher incidence of wet taps [Baby Miller]

Varieties of Epidural Anesthesia

Single-shot technique is easiest and provides the most uniform spread of anesthetic. Always begins with a negative aspiration and a test dose (3 cc of 1.5% lidocaine with 1:200,000 epi – former tells you if you’re intradural, latter if intravascular) and a 3 minute wait. If the test dose is adequate, inject the total amount in fractionated aliquots (5 cc each) as the needle location can still change.

Continuous epidural techniques involve placement of a catheter 3-5 cm beyond the needle (any longer than that and you run the risks of entry into a vein, exiting the foramen, or wrapping around a nerve root). NEVER draw the catheter back through the needle (transection). Also give a test dose and aspiration test (r/o CSF), inject in 5 cc aliquots.

Caudal blocks are epidural injections placed through the sacrococcygeal ligament and sacral hiatus (absent in 10% of patients) – pass through the ligament until you hit sacrum, then retract slightly, aim cephalad, re-advance 2 cm, inject air – if no crepitus is seen, you are likely in the caudal canal and can inject. Note that the sacral hiatus is absent in 10% of patients.

Level and Duration

By decreasing concentration and increasing volume, one can obtain greater anesthetic spread. Lumbar epidurals tend to flow cephalad due to negative intrathoracic pressure, whereas thoracic epidurals tend to stay in place. L5/S1 anesthesia is more difficult, likely due to the large fiber size. Note that in epidurals, baricity does not matter (but negative intrathoracic pressure does) in terms of levels, and body position is less important.

Chlorprocaine is used for rapid onset / short procedures. Lidocaine is intermediate, and bupivacaine/L-bupivacaine/ropivacaine have slower onset and prolonged duration. Note that tetracaine and procaine are not used because of their long latency times.

Adjuvant Medications


Epinephrine (1:200,000 i.e., 5 ucg/mL) can prolong an epidural, especially if chlorprocaine or lidocaine is used (not so much with bupivacaine). However, the mild B-stimulation may accentuate the fall in blood pressure that generally occurs with neuraxial anesthesia.


Opioids can enhance analgesia, with the degree of side effects largely related to lipid solubility. Morphine (hydrophilic/lipophobic) injected epidurally stay in place or spread rostrally, whereas fentanyl (hydrophobic/liphophilic) will be rapidly absorbed.


Sodium bicarbonate favors the non-ionized form of local anesthetics and promotes more rapid onset of epidural anesthesia.

Failure and Side Effects

If epidural anesthesia has “partially failed” i.e., anesthesia is achieved but it is inadequate and maximum doses have been approached, consider injecting small doses of chloroprocaine [Crosby et. al. Can J Anaesth 38: 136, 1991]. Note that converting to a subarachnoid block may be difficult, as anesthetic levels following an attempt at epidural are often erratic. [Mets et. al. Anesth Analg 77: 629, 1993]

Epidural hematoma has traditionally been associated with vascular trauma, but it is recognized that both epidural hematomas and abscesses can occur spontaneously.

Dural puncture significantly increases the risk of headache – epidural anesthesia can be attempted at a different level, or the procedure can be converted to a spinal.

Systemic hypotension is more delayed than that seen following spinal anesthetics, but can occur. It is rare, however, in normovolemic patients.

Absorption/intravascular injection are particularly troublesome for bupivacaine, which has known cardiovascular side effects. Epidural doses of any local anesthetic, when injected in the subarachnoid space, can lead to permanent nerve injury. If this occurs, consider irrigating the subarachnoid space with saline. This is easily recognized in an awake patient, however in a patient under general anesthesia, look for a dilated, non-reactive pupil (indicates possible migration of an epidural catheter into the subarachnoid space).

Neural injury is more likely if paresthesias occur [Baby Miller], thus injection of local anesthetics in the presence of paresthesias is contraindicated.


Major site of action for an epidural is at the nerve roots. To a lesser extent, analgesia is provided by diffusion into the subarachnoid space. The most important physiologic alteration associated with an epidural (as well as a subarachnoid) block is sympathetic blockade – T1 to T4 are the cardioaccelerator fibers which control heart rate and contractility, and their absence leaves one vulnerable to excessive vagal reflexes which can produce sinus arrest. There is debate, however, about whether or not a sympathectomy is disadvantageous in an adequately-hydrated patient. SNS block was shown to increase mucosal blood flow and bowel peristalsis, enhancing return of GI function after colon surgery [Liu et. al. Anesth 83: 757, 1995]. It also favorably alters myocardial oxygen supply, reduces ischemic events, and improves functional recovery after myocardial stunning in dogs [Rolf N et al. Anesth Analg 83: 935, 1996]. Sympathectomy-induced hypotension can reduce surgical bleeding, but if extreme can place the patient at risk for various infarctions. Epidural blocks have favorable effects on pulmonary function by preventing splinting and maintaining the ability to cough and participate in deep breathing. Note that it is theoretically possible to have a high epidural whereby the breathing apparatus is blocked but consciousness is maintained.

Cardiovascular Effects of Neuraxial Anesthesia

Neuraxial blocks results in a sympathectomy 2-6 dermatomes above the sensory block. While it is traditionally taught that the effects of a spinal anesthetic are more profound than those of an epidural anesthetic, this belief is likely related to the speed of onset – a rapid onset epidural (ex. chlorprocaine 20 cc over 3 mins) may result in equivalent cardiovascular responses.

Both arteries and veins vasodilate, however the venous effect is more pronounced, leading to substantial decreases in preload. Despite this, empiric administration of volume has not been convincingly shown to decrease hypotension following initiation of neuraxial anesthesia.

The effect on heart rate is due to three mechanisms – 1) the Bainbridge reflex, in which stimulation of right atrial stretch receptors leads to vagal afferent stimulation of the medulla and subsequent inhibition of parasympathetic activity (increasing the heart rate, or, in the case of decrease atrial pressure, lowering heart rate) 2) a direct effect on the SA node elicited by atrial stretching and 3) anesthesia of T1-4 cardioaccelerator fibers (in the setting of a high spinal).

Neuraxial Anesthesia: Cardiovascular Effects

  • Tone: arterial and venous dilation
  • Preload: decreased
  • Heart Rate: decreased (Bainbridge reflex, direct effect on SA node, T1-4 cardioacclerators)
  • Pharmacologic Treatment: ephedrine vs. epinephrine (except in parturients, in whom consider phenylephrine, but beware bradycardia and consider adding glycopyrrolate)

Post-Dural Puncture Headache

The risk of a headache after accidental dural puncture (i.e. with an epidural needle) is approximately 50%. However, keep in mind that headaches occur in 12% of all parturients who have an epidural (and 15% of parturients who don’t have an epidural).

For patients whose dura is violated intentionally (ex. spinal anesthetic, CSE, DPE), the most important modifiable risk factors have to do with needle selection – small (24 or 25 ga.) pencil point (Whitacre or Sprotte) needles should be selected. A large (22 ga.) Quincke needle can produce PDPH in 30-70% of cases, whereas a small (24-25 ga.) Whitacre or Sprotte needle will produce PDPH in only 3-5%

Risk Factors for Post-Dural Puncture Headache

  • Beveled (Quincke) needle (pencil-point needles are preferable)
  • Larger needle
  • Female gender
  • Pregnancy
  • Younger age
  • History of headache prior to the dural puncture

Miscellaneous Notes from Barash

[note: references will need to be verified, and content incorporated into the rest of this page] The site of action for central blocks is unknown, but there is some evidence that it is peripheral – Boswell’s study of a spinal anesthetic showed cortical evoked potentials during direct spinal cord stimulation but absence of SSEPs when the tibial nerve was stimulated. Note that SSEPS are maintained during epidural blocks, although the amplitude is decreased and latency increased.

Block sensitivity: SNS > pain > touch > motor (last to go). Occasionally patients will feel no pain but will still have intact touch.

Central blocks decrease MAC [158], produce sedation [154], and potentiate hypnotics [155-157]. Central blocks can cause a severe sympathectomy, leading to bradycardia and hypotension. Some studies relate the incidence of sympathectomy to block height [160, 159 poorly] while others do not [161]. Both spinal and epidural anesthesia have been shown to produce sudden, unexplained bradycardia or even asystole [170-1] (note: cardioaccelerators are from T1-4). Pre-existing 1st degree block may be a risk factor for progressing to a 2nd or 3rd degree block during spinal anesthesia. Epidural with epinephrine seems to cause more hypotension (20% drop in MAP) than epidural without epinephrine or spinal, both of which cause a 10% drop in MAP. [Tolas Acta Anaesth Scand [suppl] 23: 429, 1966]

Spinal Block

For a unilateral procedure, your alternatives are to either A) use a hypobaric solution (ex. with the operative hip up) or B) use a hyperbaric/supine combination provided that the patient is able to lie with the operated side down for at least six minutes prior to rolling supine [Martin-Salvaj G et al. Anesth Analg 79: 1107, 1994]. Note that at some point, as the analgesic agents become diluted out of dextrose, the anesthetic becomes “fixed” and position no longer matters. Studies by Pvery et al [45, 46] and Bodily et al [51] however, have shown that this can take over 60 minutes to occur. Also note that there are no true isobaric solutions, all of them are slightly hypobaric. Additionally, “isobaric” solutions are some of the most widely variable in terms of distribution but in general, isobaric solutions offer lower blocks than hyperbaric solutions, decreasing the overall risk of cardiovascular compromise.

Baricity and patient position overwhelmingly affect the block height. Minor contributors such as concentration, dose, and volume, are of almost trivial consequence. Injection site seems to matter for isobaric solutions [74] but not for hyperbaric solutions [70]. Morphologic (ex. height) and various descriptive CSF variables (ex. CSF volume) have been shown to be statistically significant but are clinically irrelevant as they are far too variable and many are difficult to assess.

Spinal blocks wear off in a cephalad to caudad direction, thus sacral levels will last longer than thoracic. Higher blocks usually wear off faster than lower blocks. Adrenergic agonists can prolong a spinal block, with the maximum does of PHE (5 mg) usually providing longer duration than the maximal dose of EPI (0.5 mg). Clonidine can also prolong blocks, even when given orally [96-98] but has been associated with increased hypotension in some studies. Agonists are most effective in tetracaine blocks, less-so with bupivacaine.

Epidural Block

Injection site matters for epidurals as well. Caudal blocks will anesthetize the sacral dermatomes / lower extremities. Lumber blocks will generally anesthetize from T6-L4 (or as much as T4-S1 if enough local anesthetic is used). Thoracic epidurals will anesthetize the thoracic dermatomes – always remember to decrease the anesthetic dose by 30-50% if the upper thoracic dermatomes are anesthetized, as the risk of cephalad spread is increased. Dose and volume are important for epidural anesthetics, while concentration is not. According to the majority of studies, position is not important. For a one-shot epidural, assume a 20 cc dose via lumbar injection will provide a mid-thoracic level block and then adjust volume as you see fit (ex. decrease volume if you only need lower dermatomes). Bupivacaine produces substantial sensory block with minimal motor block (as opposed to etidocaine, which has relatively high motor block). Epinephrine at 1:200,000 can prolong a lidocaine block, but not a bupivacaine block (mechanism unknown – possibly decreased blood flow, intrinsic analgesia provided by epinephrine, or increased volume of distribution.


Miller, RD et al. Miller’s Anesthesia, 7th edition, Churchill Livingstone: p 409, 1616-8. 2009