Article Text

Download PDFPDF

Obstetric anaesthesia: what a neurologist needs to know
  1. Mary Angela O’Neal
  1. Department of Neurology, Brigham and Women’s Hospital, Roslindale, MA 02115, USA
  1. Correspondence to Dr Mary Angela O’Neal, Department of Neurology, Brigham and Women’s Hospital, Boston, MA 02115, USA; maoneal{at}


Neurologists are often consulted to see women postpartum who are having difficulties involving the lower extremities; weakness, numbness and pain. Many of these women have received labour analgesia. Often, there is limited understanding by the neurologist of how these procedures are performed, why a neuraxial technique is chosen and their potential complications. This case-based review will explain the differences in the neuraxial procedures: epidural, spinal and combined spinal epidural; their advantages and disadvantages, why one technique might be chosen over another, contraindications as well as procedural risks.

  • neuraxial anesthesia
  • obstetrics
  • complications
View Full Text

Statistics from

Neuraxial procedures


The spinal cord ends in the conus medullaris, usually at the L1 vertebral body level, although sometimes it can be lower.1 The spinal cord is covered by three membranes: the pia mater (innermost membrane), the arachnoid mater and the dural mater (outermost membrane). The dura and arachnoid mater layers are loosely attached to form the dural sac, which contains the spinal cord and cauda equina. The subarachnoid space contains the cerebrospinal fluid (CSF), vessels and spinal nerves. The epidural space is between the dural sac and the inside of the bony spinal canal. The ligamentum flavum forms the posterior border of the epidural space at each interlaminar space.

Sensory levels required for labour analgesia

Neuraxial anaesthesia is commonly used for analgesia during labour. The level of sensory anaesthesia required depends on the targeted pain or the necessary procedure. For example, analgesia for the first stage of labour requires a block from T10 to L1, and this needs to extend to the sacral levels during the second stage of labour. Anaesthesia for a caesarean section requires a T4 sensory level of analgesia due to referred pain from peritoneal manipulation.

Description, advantages and disadvantages of each technique

The choice of procedure—spinal, epidural or combined spinal epidural—is individualised based on patient characteristics, such as need for rapid pain relief, duration of anaesthetic action and likelihood that the patient will need a more extensive block. Spinal anaesthesia is generally given by a single injection, whereas epidural anaesthesia is usually given via a catheter; a combined spinal epidural combines the two approaches. In general, all labour techniques should be performed below the conus: procedures performed at too high a level or where there is an abnormally low conus risk causing neurological injury.

The medications used for neuraxial anaesthesia are combinations of local anaesthetics, opioids, adjuvants and alpha-adrenergic agonists. The extent, duration and density of the anaesthetic block is determined primarily by the agents used, the medication dose, patient positioning and baricity (the specific gravity of the anaesthetic agent compared with CSF). A drug’s baricity determines how it will spread in the intrathecal space. For example, with the patient upright, a hyperbaric medication would be expected to fall in the subarachnoid space.

Spinal anaesthesia occurs when anaesthetic medication is injected into the subarachnoid space. This acts on the neural elements directly causing a block in motor and sensory nerves below the level injected. The extent of the block is determined by patient positioning, the type and dose of medications selected. The needle is placed below the conus medullaris at the L2 level. The technique requires first using an introducer needle that is advanced to the ligamentum flavum; then a small, 24 to 27-gauge, non-cutting needle is inserted into the introducer and moved into the subarachnoid space. Once CSF is obtained, the spinal anaesthetic can be injected. Spinal anaesthesia is generally a single shot and is commonly used for short surgical procedures such as forceps delivery, caesarean section or perineal trauma. The most common adjunct is a short-acting opiate; a longer-acting opiate such as morphine or diamorphine can supplement this for longer postoperative analgesia. Spinal analgesia provides a rapid symmetric block and requires low doses of local anaesthetics and opioids. The spinal injection is a small one-time dose; although initially the motor and sensory block is intense, it wears off very quickly, usually within hours. The duration is limited by medication half-life and the level of the block cannot be extended.

An epidural block for labour analgesia is usually placed well below the conus medullaris, most commonly at the L4–5 area; higher levels can be chosen depending on anatomy. An epidural delivers drugs outside the dura; its main effect is on nerve roots leaving the dura at the level of the epidural, and it gives less neuromuscular blockade than spinal anaesthesia. Typically, a 16 to 18-gauge needle is inserted and advanced through three sets of ligaments until it is in ligamentum flavum; the stylet is then removed and a syringe containing either air or saline is attached. With the needle tip in spinal ligaments, the contents of the syringe are difficult to inject. As the needle enters the epidural space, the resistance to injection is lost and the syringe contents are easily injected. The syringe is then removed, and a catheter is carefully threaded into the space. The major constituents of the epidural space are fat and blood vessels. Once the catheter is in place, the operator draws back on the syringe; if there is no blood return, a test dose of the anaesthetic agent is injected. This ensures the catheter is not in a vessel or unintentionally in an intrathecal location. If the operator either fails to recognise that the catheter is in the intrathecal space or gives inadequate time after the test dose of local anaesthetic, then the large volume of local anaesthetic intended for the epidural space would be injected into the subarachnoid space, which could result in a total spinal block.2 The symptoms of a total spinal block include a rapid ascending sympathetic, sensory and motor block with associated bradycardia, hypotension, dyspnoea, swallowing difficulty, respiratory arrest, hypotension and bradycardia.

An adequately spread block from T10 to S1 to cover the first stage of labour requires a long-acting anaesthetic, such as bupivacaine or ropivacaine. The block is commonly supplemented by short-acting opiates such as fentanyl or sufentanil. The epidural catheter allows medication to be administered by continuous infusion, boluses or a combination of these. The doses of medication used for epidural analgesia are approximately 10 times larger than those used for a spinal block, which is why a test injection is required after placing an epidural. In addition, as an epidural is commonly in place for 12–24 hours, the total dose of medication is high. This influences the degree of motor block and the rate of neurological recovery, which is longer than that of spinal anaesthesia.

Because epidural anaesthesia is catheter based, it allows for a longer period of pain relief but has a slower onset time, typically 15–20 min. It is not the neuraxial technique of choice for an elective caesarean section but is commonly used when already in place for an emergency caesarean section. Furthermore, due to the relatively large needle size required for catheter placement, if an unintended dural puncture occurs, there is a greater risk for developing a postdural-puncture headache.3

A combined spinal epidural is a combination of both techniques. First, the epidural needle is placed at or below L2 and then a spinal needle is inserted into the subarachnoid space and the anaesthetic injected. The operator removes the spinal needle, positions an epidural catheter and then injects a test dose of medication. This combined technique can be used to start labour analgesia as well as for an elective caesarean section. Combined spinal epidural has advantages of both techniques, with rapid onset, symmetric block and the flexibility of extending both the duration and the level of the block. The recovery time from combined spinal epidural is longer than that of a spinal epidural, due to the type and dose of local anaesthetic used. Combined spinal epidural also carries a higher risk of both postdural-puncture headache and medication errors causing a high spinal block.4

Neuraxial procedures are preferred for labour analgesia given their safety profile. Contraindications for neuraxial anaesthesia include patients who have a coagulopathy, pathology that puts them at risk for herniation and abnormal spinal cord anatomy. Also, patients who take anticoagulants or clopidogrel, and women with myasthenia gravis need special considerations.

Factors that obstetric anaesthetists consider when choosing a neuraxial technique

Case 1

A 39-year-old woman G*6 P**1, presented at 33 weeks pregnant with severe headache, nausea and vomiting associated with right-sided weakness (G*=gravida—number of pregnancies; P**=parity—number of births).

On examination, her blood pressure was 180/100 mm Hg. She was lethargic with a right visual field cut, right hemiparesis, sensory neglect of the right leg and a right Babinski sign.

Abnormal laboratory investigations included a platelet count 105×109/L (150–450), international normalised ratio of 1.1, abnormal liver function tests, low haematocrit, an elevated fibrinogen concentration and proteinuria 4+ on urinalysis.

CT scan of head without contrast showed a left frontal/parietal intracerebral haemorrhage (figure 1).

Figure 1

Head CT without contrast showing a left frontal/parietal intracerebral haemorrhage with some surrounding oedema.

This woman had severe pre-eclampsia manifesting as haemolytic anaemia, elevated liver enzymes and low platelets (HEELP syndrome) complicated by an intracranial haemorrhage.

She needs an emergency caesarean section to treat the pre-eclampsia. What are the appropriate concerns for the obstetric anaesthetist? First, she has a coagulopathy. To offer safe neuraxial analgesia, anaesthetists want platelet counts of 70–100×109/L and would not consider this procedure for patients with platelet counts below 50×109/L. They must also consider whether the platelet function is normal or abnormal as in pre-eclampsia. This patient’s platelet count is adequate for the procedure.

Next, she needs a rapid and complete block for the planned caesarean section. This would be best obtained with a spinal or a combined spinal epidural procedure. However, she has an intraparenchymal haemorrhage with mass effect, which is a contraindication for all neuraxial procedures as both spinal and combined spinal epidural could cause herniation and an epidural with an unintended dural puncture could lead to the same result. Ultimately, it was decided that general anaesthesia was the safest choice for the patient.

Case 2

The obstetric anaesthetic team was consulted for recommendations on neuraxial anaesthesia before delivery for a 29-year-old woman (G3 P1, 39 weeks pregnant) who had suffered bithalamic strokes 12 months before (figure 2).

Figure 2

Brain MRI: on the left is the DWI and on the right is the ADC sequences showing acute bithalamic strokes.

Haematology consultants recommended therapeutic anticoagulation in the setting of pregnancy and prescribed subcutaneous enoxaparin injections.

The major issue here is that her anticoagulant needs to be stopped before any neuraxial procedure. Her strokes are small and remote. Patients on either anticoagulants or clopidogrel need to stop them before considering any neuraxial procedure. The length of time necessary off anticoagulation/antiplatelet medication relates to drug half life (see table 1).5

Table 1

Recommendations based on the American Society of Regional Anesthesia and Pain Medicine(5)

Case 3

A 34-year-old woman (G1 P0, 35 weeks pregnant) was evaluated to define the safety of neuraxial anaesthesia. She had previously undergone surgery for spina bifida and a tethered cord. Her neurological impairments included weakness and numbness in legs, distal more than proximal and urinary incontinence.

Her paraparesis increases the risk of requiring either an assisted delivery or a caesarean section. Therefore, the best neuraxial technique is one in which the level of the block can be extended if necessary. Her conus medullaris likely lies abnormally low, putting her at risk for spinal anaesthesia. However, epidural anaesthesia can be placed above the level of the surgery. Her potential complications of epidural analgesia relate to surgical scarring causing difficulty locating the epidural space, an asymmetric block and difficulty placing the epidural catheter.

A MR scan of lumbar spine without gadolinium was obtained to understand her anatomy (figure 3). After consultation, it was decided to place an epidural above the level of the surgery. She presented at 38 weeks in labour. A successful epidural was placed at T12–L1 under ultrasound guidance. Due to failure to progress, a caesarean section was performed.

Figure 3

Axial and sagittal T2 lumbar MRI sequences showing postsurgical changes from L3–L5 with neural elements present to the sacrum with an associated syrinx.

In women with myasthenia gravis, the preferred technique is neuraxial anaesthesia. If local anaesthetics are used, then ropivacaine, mepivacaine, bupivacaine or lidocaine should be chosen; esters and opiates may exacerbate myasthenia and should be avoided. In addition, special care should be given to avoid a high block as that can precipitate respiratory compromise, especially in tenuous patients.

Complications of neuraxial anaesthesia that may involve the neurologist

A review of published studies showed that significant complications from neuraxial procedures were rare, at 1 per 3000 women. These complications included high neuraxial block, respiratory arrest in labour and delivery and an unrecognised spinal catheter.6 This contrasts to the relatively high incidence of obstetric neuropathies, which can complicate 1% of all deliveries.7 These obstetric neuropathies are usually compressive and relate to positioning, prolonged labour or fetal/maternal disproportion and do not relate to the anaesthetic technique.

Postdural-puncture headache

This is the most common complication of neuraxial anaesthesia. An unintended dural puncture complicates 0.7% of neuraxial procedures6 and over half of these develop a postdural-puncture headache.

  • Patient factors that increase the risk include female gender, low body weight, previous postdural-puncture headache and a history of migraine. Thus, the obstetrical population is particularly vulnerable.

  • Operator factors that increase the risk include needle size, use of a cutting needle and multiple attempts to place the epidural.

The symptoms include a bilateral headache that often radiates to the neck, sometimes with diplopia and pulsatile tinnitus. These symptoms significantly improve with lying flat. Nausea, neck stiffness and photophobia are also common.

Postdural-puncture headache is generally benign and self-limiting and usually resolves with bed rest. Subdural effusions are relatively common but subdural haematomas are rare (1 in 3900).8 In general, obstetric anaesthetists handle this complication. Neurologists are usually consulted when there are atypical features such as minimal positional change in headache severity; neck pain radiating to shoulders, arm and spine; cranial nerves symptoms and lack of response to an epidural blood patch.9

Caffeine showed benefit in a few randomised controlled trials. Women who do not improve after 24–48 hours of bedrest or who have cranial nerve involvement should be offered an epidural blood patch. This is highly effective with headache resolving in 90% of patients.10 The technique involves injecting 15–20 mL of the patient’s blood into the epidural space. Contraindications to placing an epidural blood patch are the same as those for epidural anaesthesia.

High block

Case 4

Neurology was consulted urgently for a 29-year-old woman G2P1 who developed right-sided weakness and ptosis. An lumbar epidural had been placed during labour. Following this, she developed sudden-onset right ptosis, with weakness of the right arm and right leg only, but no shortness of breath. There were no sensory findings.

On examination, her vital signs were stable. Her right pupil was slightly smaller than the left, with associated ptosis. There was a right pronator drift and subtle leg weakness especially on the right.

Urgent CT scan of head and CT cerebral and cervical angiograms were normal, as were MR scans of brain and cervical spine. She returned to her baseline over several hours.

Horner’s syndrome may follow an injury to the ipsilateral sympathetic nerves to the eye. First-order neurones arising from the posterolateral hypothalamus descend uncrossed to the intermediolateral cell column of the spinal cord at the level of C8–T2; second-order preganglionic fibres exit the spinal cord at T1 to follow the cervical sympathetic chain, where they are near the pulmonary apex. They synapse in the superior cervical ganglion at the level of the bifurcation of the common carotid artery; the postganglionic pupillomotor fibres exit the superior cervical ganglion and ascend along the internal carotid artery to the eye. The possible causes of this patient’s Horner’s syndrome include a brainstem stroke, a cervical cord injury or a direct damage from the neuraxial anaesthesia. A brainstem injury could cause a right Horner’s syndrome, but there should be other brainstem signs and the weakness would be contralateral. A cervical cord injury is possible but unlikely given the absence of predisposing factors such as hypotension, no sensory deficits and the relationship to the epidural anaesthesia. In this case, the Horner's syndrome with ipsilateral arm and leg weakness following epidural analgesia was likely due to an asymmetric high block on the right side due to the catheter being in the subarachnoid rather than the epidural space.

Spinal cord ischaemia

This complication is rare in obstetric anaesthesia and more commonly relates to abdominal aneurysm repair or major spine surgery.11 12 The anterior two-thirds of the spinal cord is supplied by the anterior spinal artery with multiple radicular feeders from the aorta, the largest of which exits from T8 to L1, the artery of Adamkiewicz. Thoracic cord infarction usually involves the artery of Adamkiewicz. Cord infarction can follow profound hypotension, hypercoagulable states, cardiac emboli and surgery or instrumentation that interferes with the vascular supply of the cord. The neurological sequelae include leg paralysis, loss of pain and temperature sensation below the level of infarct, urinary incontinence and preserved vibration and joint position sense in the lower limbs.

Haematoma/bleeding complications

Case 5

A 33-year-old woman G3P2 presented with back pain 2 days after a normal vaginal delivery. She had received an epidural block during her delivery. The epidural had been difficult to place and had required multiple attempts. She reported an electric-like pain in the low cervical/high thoracic region that radiated to her occiput and occasionally to her sacrum. She had multiple episodes of this brief pain, described as ‘putting a finger in a light socket’, most of which were precipitated by positional changes. She also reported a bifrontal headache worse with sitting or standing. She had no weakness, sensory or bowel or bladder symptoms.

On examination, her vital signs were stable and she was afebrile. There was no back tenderness to palpation. There was moderate nuchal rigidity but her neurological examination was normal.

Her laboratory investigations were all normal. Figure 4 shows her MR scan of lumbar spine.

Figure 4

Axial and sagittal T2 lumbar MRI sequences showing subacute blood in both the subdural and subarachnoid spaces. The arrows are highlighting the dark signal which is methemoglobin.

In this case, she was experiencing meningeal irritation from subarachnoid blood as well as a postdural-puncture headache, due to an unintentional traumatic dural puncture. She was neurologically normal, so required only supportive care (whereas a traumatic haematoma with significant motor deficits would have needed urgent neurosurgical intervention). Following neuraxial anaesthesia, the incidence of haematoma is 1 in 1 62 453.6 The risks are increased with antiplatelet medication, anticoagulants, trauma and an associated coagulopathy such as pregnancy-induced thrombocytopenia or the HELLP syndrome. Clinicians should suspect an epidural haematoma following neuraxial anaesthesia in patients who have unusual back pain as well as motor, sensory or sphincter deficits persisting beyond the expected duration of the local anaesthetic. MR scan of spine is the best imaging modality. In general, spinal haematomas present early, often when the epidural catheter is in place, whereas epidural abscesses present later (median 5–8 days postprocedure).13

Infection: epidural abscess, meningitis

The incidence of an epidural abscess or meningitis in the obstetric population is 1 in 62 866.3 The development of an epidural abscess is usually delayed.14 Most pathogens arise from skin flora. Risk factors for infection include immune compromise, poor adherence to sterile technique and a long duration of the catheter left in place.15 Patients generally present with fever, severe focal back pain with or without signs and symptoms of cord and root involvement.

Meningitis can also occur after neuraxial procedures; this can be either bacterial or chemical. Predisposing factors for bacterial meningitis are similar to those for epidural abscess. Streptococcus viridians is the most common pathogen.

Direct injury

Direct injury to the spinal cord occurs when there is a low-lying conus or the level of the conus is misjudged. In 20% of patients, the conus extends to the level of L2. To minimise this risk in spinal and combined spinal epidural anaesthesia, the needle is positioned below this level. A cord injury from an accidental dural puncture related to an epidural can occur if the epidural is placed in the thoracic spine or if the conus is low-lying. Trauma can also occur directly from the catheter or needle to the nerve roots or may relate to intraneural injection of the anaesthetic agent. In patients with known abnormal spinal anatomy or prior surgery, imaging is important to plan the safest location for placement of an epidural, as illustrated by Case 2.

Distinguishing a spine or root problem from an obstetric neuropathy/plexopathy

Obstetric palsies (compressive injuries related to childbirth) are much more common than complications from neuraxial anaesthesia. They are almost always demyelinating and have an excellent prognosis. The lateral femoral cutaneous nerve is the most commonly injured nerve in the lower limbs, followed by (in decreasing incidence) the femoral nerve, fibular nerve, lumbosacral plexus, sciatic and obturator nerves.7 Table 2 includes the key examination features to distinguish these neuropathies and their appropriate investigations.

Table 2

Features distinguishing obstetric neuropathy and plexopathy from spinal cord and nerve root pathology

Key Points

  • Neuraxial procedures: spinal, epidural and combined spinal epidural vary in their advantages and disadvantages.

  • Neuraxial procedures require special safety considerations in patients with a coagulopathy, increased intracranial pressure related to a brain mass, abnormal spinal anatomy or myasthenia gravis.

  • Neuraxial anaesthesia rarely causes neurological complications.

  • Neurologists need to distinguish complications related to neuraxial anaesthesia from the more common postpartum obstetric neuropathies.


View Abstract


  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Commissioned; externally peer reviewed by Niel Anderson, Auckland, New Zealand, and Rachel Collis, Cardiff, UK

  • Data sharing statement You can share any data from this paper.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.