Identifying patients with an increased bleeding tendency is essential, and several guidelines exist for performing spinal procedures in patients with an increased bleeding tendency.^97,¹²¹ Information including a medical and family history of coagulation disorders, a careful medication history to identify the use of anticoagulants, and a history of prolonged bleeding after trauma or prior surgery should be obtained from every patient. If a coagulation disorder is diagnosed and managed, spinal procedures may be considered after evaluation and disclosure of the expected risks and benefits.^181,¹⁹⁶

Antiplatelet Medication

Existing guidelines for spinal procedures in patients receiving antiplatelet and anticoagulant therapy vary by the type of medication.^96,^97,^108,¹²¹ Nonsteroidal antiinflammatory drugs (NSAIDs), including aspirin, reduce platelet aggregation, but have not been implicated in increased bleeding risk or complications in lumbar epidural injections.⁹⁶ In patients with otherwise normal coagulation status, discontinuation of these drugs is not required.

Clopidogrel (Plavix) and ticlopidine (Ticlid) inhibit adenosine diphosphate–induced platelet aggregation, but bleeding risk after spinal injection has not been studied. Based on labeling precautions, surgical literature, and interventional cardiology and radiology data, it is recommended that before spinal intervention, clopidogrel be discontinued for 7 days and ticlopidine for 14 days.^108,¹²¹

Glycoprotein IIb/IIIa receptor antagonists such as abciximab, eptifibatide, and tirofiban inhibit the binding of thrombocytes to fibrinogen and von Willebrand factor, thereby inhibiting platelet aggregation. This class of medications is an absolute contraindication to spinal procedures. Because these medications are used in the management of acute coronary syndromes, the physician is unlikely to encounter them when assessing for a spinal procedure. Discontinuation of eptifibatide and tirofiban for 8 hours before an intervention, and for 24 to 48 hours for abciximab, is recommended by guidelines for regional anesthesia.⁹⁷

Anticoagulants

Warfarin has a profound impact on coagulation as a vitamin K antagonist. Active use of warfarin results in an increased risk of spinal hematoma after lumbar puncture.⁹⁷ The use of warfarin for therapeutic anticoagulation is therefore considered a contraindication for spinal interventions. Warfarin should be discontinued for 4 to 5 days before any spinal intervention to allow the prothrombin time and international normalized ratio to return to the normal range. For invasive procedures such as spinal interventions and surgery, an international normalized ratio of 1.5 or less is considered safe in most guidelines.^97,¹⁰⁸

Low-molecular-weight heparin (LMWH) is administered either for thrombus prophylaxis or therapeutically at higher doses. After a prophylactic dose of LMWH, it is recommended that spinal procedures be postponed for 12 hours. The medication must be held at least 24 hours before the procedure when a patient is receiving a therapeutic dose of LMWH, and cannot be administered until 24 hours after the procedure.¹²⁵

Unfractionated heparin is administered either as a prophylactic or therapeutic dose. Prophylactic low-dose unfractionated heparin is not a contraindication for spinal procedures.¹⁹² Therapeutic doses of heparin should be discontinued 2 to 4 hours before any spinal procedure and a normal activated partial thromboplastin time documented before the procedure. Heparin may be restarted 1 hour after the procedure.¹²⁵

Bleeding: Diagnosis and Management

The risk for a spinal hematoma is present regardless of the patient’s hemostasis status. The outcome after spinal hematoma is highly dependent on the time interval between symptom onset and treatment with surgical decompression. The prognosis of spinal hematoma is poor without definitive treatment within 8 to 12 hours after onset of symptoms.¹⁶¹ Warning symptoms include sensory and/or motor block longer than expected, recurrence of sensory and/or motor block after initial resolution, diffuse sensory and/or motor block beyond what was expected, and unexplained spine pain with or without radicular pain.^116,¹⁶¹ These symptoms are not specific and can present immediately or be delayed several hours to days.¹¹⁶ Specific postprocedure instructions, including a contact if symptoms are encountered, are essential. Emergent magnetic resonance imaging (MRI) is required for diagnostic evaluation, followed by emergent decompressive surgery if a spinal hematoma is present.

Allergy

Review for any prior history of allergic reaction to medication is essential. Assessing the type of reaction for each medication can prove helpful. A positive history for urticaria, swelling, rashes, or breathing problems should be considered severe and potentially of the anaphylactic type. Although many “reactions” to medications are not truly allergic, it is prudent to clarify this if the patient consents to receive a similar medication.

Often the first medication administered in the procedure suite is a sedative. One case of a severe anaphylactoid reaction to midazolam has been reported,⁷⁸ although the risk for an allergic reaction to sedatives is exceedingly low.

Allergy to contrast is more common and is usually an anaphylactoid type of reaction. There are several mistaken beliefs about contrast allergy. Iodine is present in ionic and nonionic contrast agents, and many who believe they are allergic to contrast dye believe incorrectly that they have an allergy to iodine. Iodine is a naturally occurring substance found throughout the body and in table salt, and true iodine allergy has not been cited. Another misconception is that those who are allergic to shellfish are allergic to iodinated contrast. The allergy to shellfish is to the muscle protein tropomyosin, which does not correlate to an iodinated contrast agent allergy.¹¹⁴ Likewise, allergy to topical iodine solutions is not an allergy to iodine, but to other agents within the solution.¹¹⁴ Similarly, allergy to iodinated contrast dye is not actually due to iodine.⁶ Consequently a patient should not be considered allergic unless the patient has had an allergy to contrast dye.

Anaphylactic food allergies in general (including shellfish, fish, eggs, milk, and chocolate), however, do correlate with an increased risk of contrast allergy, with a 5.7% incidence, whereas persons with allergic asthma have a 7.7% incidence.¹¹⁴ Screening for atopic reactions of any kind can help determine those at higher risk of a contrast allergy.

If patients have a history of a contrast allergy, they can be premedicated with prednisone, 20 to 50 mg orally via several different regimens. Oral administration must be given, however, at least 6 hours before contrast administration. Intravenous dosing immediately before the procedure is not effective.¹¹⁴ Ranitidine, 50 mg orally 2 hours precontrast, and diphenhydramine, 25 to 50 mg orally 2 hours precontrast or intravenously 10 minutes precontrast, can also be helpful. Use of gadolinium is also considered acceptable. However, these measures do not completely prevent an anaphylactoid reaction.¹³⁴

Anesthetics can also cause an allergic reaction. The ester subgroup is metabolized by plasma pseudocholinesterase into para-aminobenzoic acid (PABA). This can lead to an allergic reaction in persons sensitive to PABA in other products (such as sunscreen). It is important to realize that those with an allergy to esters are not usually allergic to amide anesthetics, and vice versa.

An allergic reaction to corticosteroid is rare, but a case of sustained urticaria after spinal injections has been reported in which the patient skin tested positive to 21 of 26 corticosteroids.¹⁵⁴ Also, adjuvants and preservatives commonly added to glucocorticoids could potentially cause an allergic reaction.

Treatment of Allergic Reaction

Allergic reactions might not be avoidable, and preparation including basic monitoring and advanced cardiac life support training is recommended. Unfortunately, many practitioners are unaware of what is considered appropriate treatment.¹⁹⁸

Treatment with high–flow rate oxygen delivered by facemask is appropriate when oxygen saturation falls below 92%.¹⁹⁸ Fluid is administered regardless of whether hypotension is present, because allergic reactions, especially those caused by contrast agents, are worsened by dehydration.¹⁹⁸ Subcutaneous or intravenous epinephrine is indicated, although the intravenous route has a more dependable uptake, particularly when a patient becomes hypotensive.¹⁹⁸ Diphenhydramine can be given, but is contraindicated if the patient is hypotensive. Consequently it has limited use in an emergent anaphylactic reaction.⁷ Corticosteroids do not modify the acute symptoms in an anaphylactoid reaction and should not be given as the sole treatment.¹⁹⁸

Procedural Risks

Medication Interactions and Toxicity

Sedatives

The use pattern of sedatives for spinal procedures is variable. Some physicians think sedatives cloud the diagnostic value of injections, while others think they are necessary and decrease anxiety and the incidence of vasovagal reactions.⁶⁸ Their use should be judicious, in view of the obvious risk of respiratory depression from overdose. Many procedures require a responsive patient to minimize risk, and oversedation could lead to serious complications. Some patients are more susceptible to conscious sedation than others; therefore slow titration is recommended.¹²⁷

The use of sedatives requires that there are no gastric contents at risk of being aspirated. The recommendation, therefore, is that the patient should have no clear liquids for 2 hours and no solids for 6 hours before the procedure. This guideline is generally followed even when sedation is not planned. Patients who have not followed this recommendation should have their procedures delayed.⁸

Contrast Agents

Using the lowest possible dose of the nonionic form of contrast in most spinal procedures helps to minimize the risk of developing renal toxicity. In those patients with normal renal function, even in the setting of diabetes mellitus, multiple myeloma, and metformin treatment, the risk of renal toxicity is low.¹¹⁴ Minimizing the dose and ensuring precontrast hydration are recommended to decrease nephrotoxicity.⁴⁹

Various medications can interact with iodinated contrast agents. Metformin can increase the risk of renal toxicity and precipitate severe lactic acidosis in diabetic patients. It might be prudent to withhold metformin after use of iodinated contrast agents in spinal procedures as well. Combining iodinated contrast administration with any nephrotoxic medication (even NSAIDs) increases the risk of toxicity.¹³⁴ Neural, cardiac, and osmotic toxicity are also possible but more rare.¹¹⁴

In spinal procedures, 0.2 to 15 mL of contrast containing between 200 and 300 mg/mL of iodine is generally used. Keeping the iodine dose below 3 g minimizes toxicity.¹¹⁴ Gadolinium can be used in place of iodinated contrast and has less toxicity.

Corticosteroids

Corticosteroids are thought to relieve pain by reducing inflammation and by blocking transmission of nociceptive C-fiber input. They have been shown to directly inhibit the activity of phospholipase A2 within the inflammatory cascade.⁹³ In addition to their antiinflammatory actions, corticosteroids also have another pain-relieving mechanism of action via nerve membrane stabilization through inhibition of ectopic impulses, ion conductance, C-fiber transmission, and hyperpolarized spinal neurons.^57,^88,^89,¹⁰⁴

Local tissue toxicity has been demonstrated with corticosteroids. Dexamethasone was shown to decrease proteoglycan concentration in human chondrocytes.¹⁸³ Although there had been some concern about neural toxicity with spinal administration, no concrete evidence of neurotoxicity from intrathecally administered corticosteroids has been reported.⁹⁴

Corticosteroids have systemic side effects that should be considered when assessing risk. A single epidural steroid injection (ESI) can cause adrenal suppression and hypothalamic-pituitary-adrenal axis imbalances for 1 to 5 weeks.¹³⁶ Complete adrenal suppression is usually limited to 4 to 7 days, but the level of incomplete suppression can be similar to that of low-dose oral corticosteroids. Some individuals are more susceptible to suppression and long-term effects, but clinical prediction of these individuals is not possible.⁶⁹ An ESI precipitates more prolonged suppression than does an intraarticular injection.²⁰⁹ Intraarticular injection studies have shown that the duration of both systemic and local effects increased with decreased corticosteroid solubility. The degree of cortisol suppression was dependent on the number of joints injected, independent of the corticosteroid dose.¹⁰

Specific guidelines for the acceptable cumulative dose of spinal corticosteroids do not exist. Many physicians advise no more than three ESIs per year or a 6-month limit of 5 mg per kilogram of body weight.¹⁸² There are no published data, however, of an evidence-based definitive guideline. Nonetheless, judicious use is recommended because the additive local effect on the spinal structures is unknown, and the systemic effects of prolonged adrenal suppression are detrimental.

Endocrine

Patients with diabetes mellitus have an increased risk from the effects of adrenocortical suppression. After an ESI, persons with diabetes can have elevated blood glucose levels particularly during the first 4 to 7 days postinjection. Some studies have found the severity of this glucose increase related to the hemoglobin A1c level at the time of injection.²⁰⁹ Blood glucose levels can exceed 400 mg/dL, with an attendant risk of ketoacidosis.

Cardiovascular

Corticosteroids increase cardiac output and vascular tone, resulting in increased blood pressure. Another effect can be increased demand on the myocardium.²⁰⁹

Immune System

The general infection risk is elevated in all patients after glucocorticoid injection, secondary to immunosuppression. An almost immediate suppression of T lymphocytes occurs.²⁰⁹ Newly acquired viral infections and reactivation of latent herpes virus are more likely to occur after glucocorticoid administration. Any existing infection is a contraindication to injection of glucocorticoids. Local infection at the site of proposed injection is a contraindication to injection, even when glucocorticoids are not being administered.

Preexisting immunosuppression in a patient requires careful assessment. All patients with diabetes mellitus are relatively immunosuppressed, and this varies with their blood glucose levels. The patient should also be screened for other causes of immunosuppression in the medical history.

Skeletal System

Long-term effects such as osteoporosis can occur with ESIs. No large clinical studies have assessed this risk. However, the fact that systemic absorption can approximate the effects of low oral dosing for at least 30 days suggests that bone density loss is likely after ESIs.¹⁸²

Gastrointestinal System

Glucocorticoids potentiate the risk of gastrointestinal bleeding, especially when used in conjunction with NSAIDs. Although no data clearly indicate that patients have to discontinue NSAIDs before spinal procedures because of the risk for bleeding at the site of injection,⁹⁶ several studies have reported approximately a twofold increase in gastrointestinal bleeding risk for those taking NSAIDs and glucocorticoids combined compared with NSAIDs alone.^145,¹⁹⁵ This may provide enough reason to withhold NSAIDs when performing spinal procedures with glucocorticoids.

Corticosteroid Particles

The particulate size of corticosteroids is another property that can contribute to adverse events during epidural injection.¹⁶⁵ Accidental intraarterial injection of particulates can cause serious harm in the cervical, thoracic, and even the lumbosacral spine by causing a spinal cord infarction.¹⁸² The cervical transforaminal epidural approach with particulate corticosteroid has been associated with serious neurologic compromise including death, stroke, and quadriplegia in at least 70 cases.¹⁶⁵

All corticosteroid preparations used for epidural injection are particulate except dexamethasone and betamethasone sodium phosphate.²¹ Dexamethasone has not been implicated in any of the embolic events associated with epidural injection. Although no large-scale studies have compared the efficacy of different corticosteroids, one smaller study suggests some slightly longer term effects with particulate corticosteroid in the cervical spine compared with dexamethasone.⁶²

Anesthetics

Anesthetics inhibit migration of sodium ions into the axon to slow conduction, and reduce the potassium conductance to slow repolarization. Anesthetics have known systemic toxicity, which limits the total dose (Table 25-1). There are two chemical subgroups of anesthetics, the amides and the esters. The amides are metabolized in the liver and excreted in the urine, whereas the esters are metabolized by plasma pseudocholinesterase into PABA.^54,¹⁴⁸ Systemic toxicity for all anesthetics affects the central nervous system (CNS). The initial symptoms frequently are uneasiness, tinnitus, metallic taste, and visual auras. These symptoms can progress to shivering and then to seizures. The management of CNS toxicity is airway protection and benzodiazepine administration for control of seizures.¹⁴⁸

Table 25-1 Anesthetic Pharmacokinetics

Cardiac effects of anesthetic toxicity are manifested through decreased cardiac conduction and atrioventricular block. Bupivacaine is 16 times more potent for cardiac QRS prolongation than is lidocaine. Bupivacaine also has a slow “on” and slow “off” binding to the sodium channels, which makes its duration of action longer for the anesthetic affect, but also for its toxicity. Cardiac toxicity with bupivacaine often precedes the CNS toxicity, and once present is rarely reversed, with a high fatality rate despite adequate resuscitation efforts.¹⁴⁸

Epidural anesthetics have been shown to locally inhibit platelets, fibrinolysis, and leukocyte function. It has been proposed that reduced granulocyte migration and metabolic activation at surgical sites infiltrated with lidocaine might be a mechanism of decreased pain postsurgically. This could also be a mechanism for the duration of pain relief exceeding the anesthetic action in spinal procedures.¹⁴⁸ Lidocaine and bupivacaine have been shown to have an antiinflammatory effect on nucleus pulposus–induced nerve injury, and to increase intraradicular blood flow in an animal nerve root compression model.^206,²⁰⁷

Infection

Prevention of infection starts with the proper preparation, including a sterile procedural field and maintaining sterile equipment and medications. Risk of infection cannot be completely eliminated and might occur with any spinal injection. A 1% to 2% infection rate is reported for all spinal injections, but only a 0.01% to 0.1% rate for severe infections.⁶⁵ The infection can spread through Batson’s plexus. Malposition of the needle, such as into the pelvic cavity, could cause infection with enteric organisms.²⁰³

Needle Malpositioning

Placement of the spinal needle outside the target area can result in deleterious consequences. Dural puncture, disk entry, intraabdominal puncture, pelvic puncture, pneumothorax, intravascular injection, nerve trauma, and spinal cord trauma are all possible with spinal interventions. They are best avoided with proper use of fluoroscopy and radiopaque contrast.⁸⁵

Fluoroscopy

The use of fluoroscopy is essential for spinal procedures. Radiation exposure is inherent and must be considered as part of the risk for both the patient and the interventionalist. Radiation doses to the interventionalist are approximately 0.39 mrem at the glasses per 100 cases, with a mean of 15 seconds per case.³⁸ This is well within radiation exposure safety guidelines, but protective lead aprons, thyroid shields, and glasses are recommended along with increasing distance from the x-ray beam.³⁸ Patients have higher levels of exposure per procedure, and consent for the procedure should include radiation exposure.⁷¹ Pregnancy is an absolute contraindication to fluoroscopy. The radiation risk should be taken even more seriously in younger patients.

Epidural Steroid Injections

General

Background

The concept of injecting medication into the epidural space has been around since the early 1950s.¹²³ Initial reports in the United States from the 1960s described treatment for sciatica using the caudal and interlaminar (IL) approaches.^20,⁸³ By the mid 1970s and 1980s, nerve root injections were described for the treatment of radicular pain. The transforaminal (TF) route of administering medication to the epidural space has become more popular in the past decade. This was largely prompted by the introduction of fluoroscopic guidance and reviews reporting that epidural corticosteroids administered by the conventional routes were not as effective as claimed in uncontrolled studies.^27,^109,¹¹³ Studies by Derby et al.⁵⁶ and others in the 1990s described outcome improvements with prospective studies evaluating the TF route for radicular pain.¹³² The last decade has been characterized by a more detailed look at the efficacy of epidural steroids, with a number of reviews conducted. The proportion of these published studies that were randomized controlled trials, however, remains very small. The findings of many of these studies are described below.

The usefulness of cervical, thoracic, and lumbar epidural injections remains controversial. Underlying this controversy is the dramatic increase in lumbosacral injections in the Medicare population from 1994 to 2001. Friedly et al.⁷⁷ reported that less than half of these epidural injections were performed for sciatica or radiculopathy, where the greatest evidence of benefit is available.

Epidural Anatomy

The epidural space is a tissue plane between the dura mater and periosteum and ligaments within the vertebral canal. It is contained anteriorly by the posterior longitudinal ligament and the vertebral bodies, and posteriorly by the laminae and ligamentum flavum. Its lateral borders are the pedicles and intervertebral foramina (Figure 25-1).

FIGURE 25-1 Lumbar spinal anatomy. all, Anterior longitudinal ligament; altif, anterior layer thoracolumbar fascia; dr, dorsal rami; ds, dural sac; esa, erector spinae aponeurois; grc, grey rami communicantes; i, intermediate branch; IL, illiocostalis lumborum; IVD, intervertebral disc; l, lateral branch; LT, longissimus thoracis; m, medial branch; M, multifidi; p, pia; pll, posterior longitudinal ligament; pltif, posterior layer thoracolumbar fascia; PM, psoas major; QL, quadratus lumborum; svn, sinovertebral nerve; st, sympathetic trunk; VB, vertebral body; vr, ventral rami; zj, zygapophysial joint.

The epidural space contains the spinal nerve roots and their dural sleeves, the internal vertebral venous plexus, loose areolar tissue, segmental blood supply, and lymphatics. The epidural veins form an arcuate pattern, positioned laterally at the level of each vertebral body. This is an important consideration for ESIs because venous puncture is more likely to occur laterally than with midline approaches.¹⁵⁸

Key anatomic features of the cervical spine include a thin ligamentum flavum (unfused in the midline in approximately half of individuals), absence of the interspinous ligament, and a small posterior epidural space (distance between the ligamentum flavum and dura mater).^95,¹⁵⁸ At C6–C7 and C7–T1 the epidural space has a mean width of 3 mm (1 to 4 mm).⁵

The laminae and spinous processes overlap in the midthoracic region, making a midline approach extremely difficult. The posterior epidural space is about 2 mm in the upper thoracic region and increases to about 5 to 6 mm in the lumbar region.

The artery of Adamkiewicz is a concern for thoracic and lumbar interventional procedures. It is the largest radiculomedullary artery and the major supplier of the anterior spinal artery in the lumbar region. The artery enters the spinal canal through a single intervertebral foramen in 85% of individuals between T9 and L2,^53,⁹⁸ and is located 63% of the time on the left side.²⁹ This structure has been implicated in paraplegia after lumbar nerve root block.⁹⁸

Pathophysiology

The primary cause of radicular pain is attributed to the inflammatory effect on the nerve from the nearby disk herniation. During surgery and on myelography, nerve roots appear swollen in patients with radiculopathy.^22,¹⁶³ A herniated nucleus pulposus induces marked inflammatory responses in the dura, nerve roots, and the spinal cord with high levels of phospholipase A2 activity.¹⁶⁴ The level of phospholipase A2 activity in extracts from human disk herniations was found to be 20 to 10,000 times more than that in any other human source.¹⁶⁴ Herniated disk specimens have also demonstrated increased levels of matrix metalloproteinase activity, nitric oxide, prostaglandin E₂, and interleukin-6.^82,¹⁰⁶

Injectant

Unfortunately, there are no standardized practices for administering an ESI (Table 25-2). Across disciplines and institutions the content and volume of the injectant vary significantly. A Corticosteroid–local anesthetic mixture is the most common medication used for ESIs in both academic institutions and private practices. The results for other medications vary, with a minority of centers adding opioids clonidine, or using local anesthetic alone for their injections.⁵¹

Table 25-2 Percentage of Institutions Using Given Medication for Epidural Steroid Injection by Spinal Region

The volume of the injectant varies based on the approach used. In cervical and thoracic epidural injections, a total of up to 3 to 5 mL might be used for ESIs in which the IL approach is used. In cervical and thoracic TF ESIs, however, clinicians generally use a maximum total volume of 1.5 to 2 mL. The volume used for lumbar ESIs is slightly greater, generally not exceeding 3 to 4 mL for TF ESIs, 6 to 10 mL for IL ESIs, and 20 mL for a caudal approach.⁴⁸

General Indications

ESIs have been used to treat a variety of spinal disorders. They are primarily and most effectively used to treat radicular symptoms. Currently, supportive evidence is lacking for the utility of ESIs in conditions such as postlaminectomy syndrome, degenerative disk disease, annular tears, spondylosis, spondylolisthesis, and vertebral fractures.

Complications

These procedures are not without risks, and some of the complications can be catastrophic, although infrequent. Transient complications encountered during epidural injections include insomnia, nonpositional headaches, increased back pain, facial flushing, vasovagal reactions, nausea, and increased leg pain.³⁶ Reported complications with lumbar ESIs include infection, hematoma, intravascular injection of medication, direct nerve trauma, subdural injection of medication, air embolism, disk entry, urinary retention, radiation exposure, and hypersensitivity reactions.³⁶ Cervical TF or IL injections are associated with relatively frequent minor adverse events (5% to 20%); however, serious adverse events are very uncommon (<1%).⁴⁶ In a 2008 review of complications of lumbar IL and TF ESI, Goodman et al.⁸⁵ concluded that most if not all serious adverse events can be avoided by careful technique, sterile precautions, and a thorough understanding of the relevant anatomy and fluoroscopic contrast patterns.

Approaches, Evidence, and Efficacy

Various technical approaches are available to access the cervical, thoracic, and lumbosacral epidural space. Despite the body of evidence that supports the use of fluoroscopy and the utility of the TF approach, the technical approach for ESIs varies widely. The three most common approaches include IL, TF, and caudal. The following technique descriptions are not intended to serve as a stand-alone reference for conducting the procedures.

In a 2002 survey of academic and private practices, a significant variance in the use of fluoroscopy for spinal procedures was noted. Private practices used fluoroscopy with more regularity than academic institutions, most notably in the cervical region where 73% of private practices and only 39% of academic institutions polled used fluoroscopy.⁵¹

Thoracic Epidural Steroid Injection

A paucity of literature on thoracic epidurals is available. This is most likely due to the significantly smaller incidence of acute herniated disks in the more stable thoracic spine. The incidence of thoracic disk injuries is 1 in 1 million persons per year, and these injuries account for only 0.25% to 0.75% of all disk herniations.⁹ IL and TF approaches are described for this region. Often a cervical technique is used for the upper thoracic region, and a lumbar technique is used for the lower thoracic region.

Access to the epidural space in the thoracic spine is somewhat more difficult secondary to the unique anatomic properties of the thoracic region. These include the closer proximity of the spinal cord and a narrower epidural space than the lumbar region. The overlapping laminae and spinous processes in this region limit midline access to the epidural space. It is also difficult to view the IL space on radiographs or using fluoroscopy. Botwin et al.³⁵ describe an overall prevalence of complications per injection of 20.5%. All were transient and resolved without any residual morbidity.

Interlaminar Versus Transforaminal Technique

Injectant Flow

One argument for using the TF approach in preference to the IL approach is the ability to deliver the medication directly to the ventral epidural space. Injectant within the epidural space will tend to flow in the direction of least resistance. Studies evaluating contrast patterns generally support this concept, although some conclusions are conflicting.

Kim et al.¹¹¹ conducted a recent analysis of epidurography contrast patterns for cervical IL epidural injections using the midline approach. The rate of ventral epidural spread from this posterior approach was 56.7% with 1 mL of injectant and 90% with 2 mL. Botwin et al.³⁷ evaluated the contrast patterns in lumbar IL injections in 2004 and found that approximately one third revealed ventral contrast flow and 16% were bilateral. In contrast, Manchikanti et al.¹³⁹ evaluated lumbar TF ESI flow patterns and demonstrated ventral epidural filling in 88% of the procedures and nerve root sleeve filling in 97% of procedures. In a prospective evaluation of contrast flow patterns with fluoroscopically guided lumbar ESI via the lateral parasagittal IL approach versus the TF epidural approach, Candido et al.⁴³ found the parasagittal IL approach superior to the TF approach for placing contrast into the ventral epidural space, with reduced fluoroscopy times and improved spread.

Effectiveness

The effectiveness of TF versus IL ESIs has been compared, and the results are mixed. Schaufele et al.¹⁶⁶ evaluated the effectiveness of TF versus IL ESI for the treatment of symptomatic lumbar disk herniations and found TF injections resulted in better short-term pain improvement and fewer long-term surgical interventions than the IL ESI. A literature review on the management of cervical radiculitis with IL versus TF ESI concluded that there was insufficient evidence supported by prospective studies to determine whether one technique was more effective or safer than the other.¹⁰⁰ Ackerman and Ahmad⁴ compared TF, IL, and caudal approaches for the treatment of lumbar disk herniations and found that pain relief was significantly more effective with TF injections.⁴

Interlaminar

Technique

The IL technique achieves access to the epidural space using the paramedian or midline approaches. The route of penetration from superficial to deep includes the skin, subcutaneous tissue, paraspinal muscles (paramedian approach) or interspinous ligament (midline approach), ligamentum flavum, and then entry into the epidural space. The most common technique includes a “loss of resistance approach” used to identify entry into the epidural space. Fluoroscopy with contrast is used to confirm the location (Figure 25-2).

FIGURE 25-2 Lumbar interlaminar epidural steroid injection. A, Anteroposterior view. B, Lateral view.

Evidence

The evidence is moderate for short- and long-term relief in managing cervical radiculopathy. In managing lumbar radiculopathy, the evidence is strong for short-term relief and limited for long-term relief.^2,³³ The evidence is indeterminate in the management of neck pain, low back pain, and lumbar spinal stenosis.

For the treatment of diskogenic pain without radiculitis, the evidence is lacking (level III) for short- and long-term relief.¹⁹

Transforaminal

Technique

The lumbar TF approach is considered to be more specific and allows for the smallest aliquot of injectant to be delivered to the site of pathology. The technique involves using fluoroscopy to identify a subpedicular target within the intervertebral foramen. The needle is advanced to achieve a perineural location in the “safe” triangle that lies just below the pedicle and above the dural sleeve of the target nerve root (Figure 25-3). Precautions are taken to avoid advancing the needle tip medial to the “six-o’clock” position of the pedicle in the anteroposterior view (Figure 25-4).

FIGURE 25-3 Lumbar epidural “safe” triangle.

FIGURE 25-4 Lumbar transforaminal epidural steroid injection. A, Anteroposterior view. B, Lateral view.

A slightly different technique is used to deliver medication to the S1 nerve root in its foramen. The approach is analogous to that of the subpedicular approach at lumbar levels. The target point for an S1 TF injection is on the caudal border of the S1 pedicle, just dorsal to the internal opening of the S1 anterior sacral foramen. The needle is inserted into the posterior sacral foramen just short of the floor of the sacral canal.²⁵

The technique for cervical TF injections is similar to that for lumbar TF injections. The correct oblique view of the target foramen is obtained with fluoroscopy. An entry point overlying the posterior half of the target foramen is used for initial needle placement. Care is taken to ensure that the tip of the needle is positioned over the anterior half of the superior articular process to ensure that it is not prematurely inserted too far into the foramen. The needle is advanced to the superior articular process and then readjusted to enter the foramen tangential to its posterior wall, opposite the equator of the foramen. Accurate placement is confirmed in multiple planes with radiopaque contrast and fluoroscopy.¹⁵⁹

Selective Nerve Root Block

Epidural injection literature uses the terms selective nerve root block (SNRB) and TF injection interchangeably. The term SNRB implies selectivity of the TF approach, enabling the interventionalist to implicate a specific nerve root as a pain generator by anesthetizing it. It has recently been demonstrated that L4 and L5 SNRBs are often nonselective or at least only partially selective.¹⁹³ Furman et al.⁸¹ observed that with as little as 0.5 mL of contrast, 30% of lumbosacral TF ESIs were no longer selective for the specified root level. The selective nature of the nerve root block applies to where the injectant travels, and the literature demonstrates that this might not be confined to a specific nerve root. The use of the term SNRB should consider these observations.

Evidence

Although the evidence in the literature is often conflicting, two recent randomized controlled trials support the efficacy of TF ESI. The first is a study by Riew et al.¹⁶⁰ that considered avoidance of surgery as the outcome measure, comparing patients treated with lumbar TF injections of bupivacaine plus betamethasone with those treated with bupivacaine alone. Significantly fewer (29%) of the patients treated with corticosteroid preparation than those treated with bupivacaine alone (67%) required surgery. The second study compared outcomes of patients treated with lumbar TF injections of corticosteroids with those treated with paraspinal injections of saline. At 12 months’ follow-up, 84% in the TF ESI group reported greater than 50% reduction in pain scores, compared with 48% treated with paraspinal injections.¹⁸⁶

The evidence for cervical TF ESI in managing cervical nerve root pain is moderate for short- and long-term improvement. For lumbar TF ESI, the evidence for managing lumbar nerve root pain is strong for short-term and moderate for long-term improvement. The evidence is limited in managing lumbar radicular pain in post–lumbar laminectomy syndrome. The evidence is indeterminate in managing axial low back pain, axial neck pain, and lumbar disk extrusions.² In a systematic review of TF ESI for low back and lower limb pain, the evidence for TF lumbar ESI is Level II-1 for short-term relief and Level II-2 for long-term improvement (Table 25-3).⁴⁰

Table 25-3 Evidence of Epidural Steroid Injection Efficacy

Caudal

Technique

Caudal ESIs have the least risk for dural puncture, and access is relatively easily to achieve. Because the access point for this procedure is more distal to the presumed in most cases, larger volumes of injectant are usually instilled typically to treat the pathologic lesion at L4–L5 or L5–S1 levels. Patients are placed in the prone position, and the caudal space is identified using palpation of the sacral hiatus and fluoroscopy. The spinal needle is introduced through the sacral hiatus, and accurate needle placement is ensured using contrast and fluoroscopy (Figure 25-5). Given that the dural sac typically ends at the S2 level, to avoid dural puncture the needle should not be advanced higher than S3.

FIGURE 25-5 Caudal epidural injection. A, Lateral view. B, Anteroposterior view.

Evidence

The evidence for caudal ESIs in managing chronic low back and radicular pain is strong for short-term relief and moderate for long-term relief, whereas the evidence for post–lumbar laminectomy syndrome and spinal stenosis evidence is limited.² A recent randomized equivalence trial found that between 79% and 91% of patients had significant pain relief and improvement in functional status with caudal injections in managing chronic low back pain for disk herniation with radiculitis.¹³⁸ Preliminary findings of a randomized, double-blind equivalence trial investigating patients with spinal stenosis who had chronic function-limiting low back and lower limb pain, showed that fluoroscopically guided caudal ESIs were effective in approximately 60% of subjects.¹⁴⁰ In a prospective cohort study of patients with bilateral radicular pain from spinal stenosis, caudal ESIs were found to reduce radicular pain and improve standing and walking tolerance.³⁴ The results of this study are limited by the lack of a placebo group. Other studies concluded that the effectiveness of caudal ESIs in managing lumbar radiculopathy was moderate (see Table 25-3).¹⁷³

Conclusion

Most studies collectively demonstrate ESIs in the lumbar spine to be effective in providing a more rapid resolution of acute to subacute radicular leg pain caused by a herniated nucleus pulposus. Evidence that ESIs have any beneficial effect, however, is limited in patients with acute axial low back pain or in those with chronic axial low back or radicular pain. Furthermore, studies have failed to demonstrate that ESIs reduce long-term pain or obviate the need for surgery.¹⁷³ The success rates for cervical epidural injections for symptomatic cervical disk herniations appear to be very similar to prior studies of lumbar epidural injections for symptomatic disk herniations.¹²⁴

It appears that ESIs can provide diagnostic and therapeutic benefits with certain limitations. They might be considered a diagnostic tool to determine whether factors within the epidural space are contributing to the pain syndrome. They may be considered for therapeutic effect when other noninterventional management has failed. ESIs should be considered earlier in the treatment algorithm to facilitate rehabilitation management in certain conditions such as acute radiculitis, or when oral medications cannot be used or are ineffective. ESIs should complement rehabilitation management in all cases and should not be considered a substitute for a well-directed and well-designed spinal rehabilitation program.

Summary

See Table 25-3.

• ESI can provide both diagnostic and therapeutic effects.

• ESI is most effective for treating radicular pain from a disk herniation in the lumbar and cervical spine.

• Currently studies have failed to demonstrate that ESI effectively reduces chronic axial back pain or degenerative conditions of the spine.

• ESI should be used in combination with a rehabilitation management program.

Zygapophysial Joint Interventions

General

Background

The zygapophysial joint is a known pain source with distinct innervation and pain referral maps in the cervical, thoracic, and lumbar spine.^67,^79,⁸⁰ Although the anatomy is similar throughout the spine, the prevalence of zygapophysial joint–mediated pain and the utility of zygapophysial joint procedures differ between spinal regions.

The prevalence of cervical zygapophysial joint pain has been studied in patients with whiplash injuries and was found to be 54% to 60% using double-blind dual or triple medial branch blocks (MBBs).^16,¹³⁰ Thoracic spine prevalence has been cited at 48%.¹³⁷ In the lumbar spine, the incidence of zygapophysial joint pain is between 15% and 40% in patients with chronic low back pain.^169,¹⁷¹ In younger injured workers with chronic back pain, the prevalence of zygapophysial joint pain is approximately 15%.¹⁷¹ The prevalence increases with age, reaching up to 40% in noninjured individuals with a median age of 59 years.¹⁷¹

Anatomy

Understanding zygapophysial joint anatomy is key to understanding the procedures targeting this structure. It is a synovial joint composed of the inferior articular process of the vertebra above and the superior articular process of the vertebra below with articular cartilage and a capsule. In the cervical spine, a small meniscus is also present.²⁰⁸ Innervation is via the medial branches of the dorsal rami.

The cervical zygapophysial joint innervation has been described by Bogduk.²⁴ For C3–T1, the medial branches of the primary dorsal rami of the corresponding joint travel along the waist of the articular pillars to innervate the corresponding joint (e.g., C5–C6 zygapophysial joint is innervated by C5 and C6 medial branches). The exceptions include the C2–C3 joint, which is innervated by the third occipital nerve (TON), and C0–C2 articulations, which are innervated by branches of the C1 and C2 ventral rami.

In the lumbar spine, as the medial branch courses inferiorly, it lies at the junction of the superior articular process and transverse process under the mamilloaccessory ligament of the vertebra below (e.g., the L3 medial branch lies on the L4 vertebra). Each zygapophysial joint is innervated by two medial branches, one from the level above and one from the corresponding level (e.g., the L4–L5 joint is innervated by the L3 and L4 medial branches) (Figure 25-6). The exception is L5–S1, which is innervated by the L4 medial branch and the L5 dorsal primary ramus, which lies at the junction of the sacral ala and the S1 superior articular process.²⁸

FIGURE 25-6 Lumbar zygapophysial anatomy and innervation. DPR, Dorsal primary ramus; I, inferior articular process; MAL, mamilloaccessory ligament; NR, nerve root; S, superior articular process; TP, transverse process.

Indications and Principles

No history or physical examination findings, including rotation plus extension, have been shown to predict placebo-controlled or comparative block–established cervical, thoracic, or lumbar zygapophysial joint pain.¹⁰⁵ Also, no imaging studies can provide a diagnosis; consequently the first role of zygapophysial injections is for diagnostic purposes. Zygapophysial joint injections are appropriate only after adequate noninterventional care, including medications, a physical therapy program, and patient education, has failed.

Complications

General risks applicable to all spinal procedures have been described previously in this chapter. Zygapophysial joint injections have common risks of temporary increased pain, vasovagal reaction, and bruising. Less common risks include infection, bleeding, and intravascular injection.

Severe complications of zygapophysial joint injections are rare. Using multiplanar fluoroscopy with little or no sedation nearly eliminates the risk for nerve root injury or entry into the spinal canal, although secondary lumbar radiculopathy, paraspinal abscess, epidural abscess, infective diskitis, and septic zygapophysial joint arthritis have been described.^172,¹⁷⁹ There are no case reports of complications from MBBs.

Zygapophysial Intraarticular Injection

Technique

The target joint is visualized fluoroscopically. The exact view varies between spinal levels, between individuals, and even from side to side. The needle is advanced to lie at the joint line and within the capsule. The joint capsule is redundant, and capsular flow can be obtained without visualization of the joint line. Confirmation of placement must be made using the smallest amount of contrast adequate to confirm placement within the joint space (0.1 to 0.3 mL). Next a preparation of corticosteroid mixed with anesthetic is administered into the joint until capsular resistance is encountered (0.5 to 1.75 mL). Care is taken not to overdistend and disrupt the capsule.²⁵

Evidence

No studies have compared the diagnostic utility of cervical intraarticular injection with that of cervical MBBs, but there is a theoretical risk of needle placement or capsular leakage into or near the vertebral foramen with intraarticular injection, and consequently less specificity.¹¹² Intraarticular injection with corticosteroid and anesthetic is routinely performed for diagnostic purposes, but has not been validated. In uncontrolled studies, intraarticular injection has shown some therapeutic efficacy.^110,¹⁷⁵ Only one randomized controlled trial (RCT) of therapeutic cervical zygapophysial intraarticular injections has been published, and interestingly, corticosteroid was not shown to be superior to the injection of bupivacaine alone.¹⁵

Data regarding thoracic intraarticular injections are not sufficient to make any conclusions on their efficacy.¹² In a recent review of lumbar zygapophysial joint injection procedures, there were no intraarticular studies of sufficient quality to be entered into the analysis.^44,⁵⁵

Diagnostic Zygapophysial Joint Intervention

Background

MBBs are considered the primary means to test a hypothesis that pain is zygapophysial joint mediated by delivering anesthetic specifically to the medial branches. Modified dual comparative blocks are considered the criterion standard for diagnosing zygapophysial joint–mediated pain, giving the highest positive predictive value (PPV) and the lowest false-positive (FP) rate. This technique uses two separate injections with anesthetics of differing lengths of action. A positive test requires that a minimum length of relief is obtained with each (e.g., 1 to 2 hours of relief with lidocaine and ≥3 hours’ relief with bupivacaine), in addition to a high level of relief (>80%).⁶³ Diagnostic blocks are generally a precursor to more definitive treatment with medial branch radiofrequency neurotomy.

Technique

After proper preparation the target medial branch at its osseous landmarks is viewed fluoroscopically. The spinal needle is advanced in the oblique view (Figure 25-7, A), and placement is confirmed in multiple planes with radiopaque contrast and fluoroscopy (Figure 25-7, B). Vascular uptake is ruled out, because limitation to the procedure exists if venous uptake occurs (even if needle position is corrected with lack of subsequent venous uptake), incurring an associated 50% false-negative rate.¹⁰⁷ Once placement is confirmed and without vascular uptake, a small amount (0.3 to 0.5 mL) of anesthetic is injected.

FIGURE 25-7 Lumbar zygapophysial medial branch blocks. A, Oblique view. B, Anteroposterior view.

In the cervical spine, anesthetic blockade of the TON is used for diagnosing C2–C3 joint pain, or anesthetic blockade of the two corresponding medial branches is used for diagnosng C3–C4 to C7–T1 pain. No validated MBB technique exists for the thoracic spine. Current technique is based on principles from validated cervical and lumbar techniques.¹² In the lumbar spine, anesthetic blockade of the two corresponding medial branches (or the L5 primary dorsal ramus) is performed for each joint.²⁵

Postprocedure provocative maneuvers are essential to determine the result of diagnostic injections. Patients must rate their pain at regular intervals in a pain diary. A positive response is considered when the duration of relief is commensurate with the anesthetic used and there is greater than 80% pain relief.

Evidence

In the cervical spine, with the use of comparative blocks as the criterion standard, the PPV of a single cervical MBB has been cited at 60% to 73%, with an FP rate of 27% to 40%.^13,^14,^129,¹³⁷ With placebo injections used as the criterion standard, the PPV increases dramatically for comparative MBBs to 81%, with a FP occurrence rate of 19%.¹²⁹

The FP rate in the thoracic spine is also high. By using comparative or modified comparative blocks and by requiring at least 75% pain relief, the single block FP rate is between 42% and 58%.¹²

The specificity of lumbar MBBs for achieving the targeted nerve has been established using computed tomography (CT) confirmation of contrast flow postprocedure. With fluoroscopically guided precise needle position, injected contrast incorporates the lumbar medial branches and the L5 dorsal ramus without spread to alternative structures that are also potential pain generators.⁶⁵

Quantifying the amount and length of relief is essential in evaluating lumbar zygapophysial joint blocks. Schwarzer et al.¹⁷⁰ found that the PPV was 31%, with a 69% occurrence of FP responses without the use of standard modified comparative blocks. Dreyfuss,⁶⁵ however, found that with the use of modified comparative blocks, the PPV was 68%, with a 32% occurrence of FP responses. Requiring a duration of improvement coincident with the type of anesthetic used and requiring high-grade relief substantially improves the PPV and decreases the occurrence of FP responses.

Sacroiliac Joint Interventions

Background

The sacroiliac joint (SIJ) is a synovial joint with documented, clinical nociceptive capacity. Nociception from the articular cartilage, as well as from the posterior joint capsule and ligaments, has been supported.¹⁶⁸ The prevalence of SIJ-mediated pain in those with suspected SIJ pain by physical examination is only about 15%.^135,^141,¹⁶⁸ The incidence of the SIJ as the primary pain generator of lumbopelvic pain is lower than that of radicular or zygapophysial origin. The incidence, however, might increase after trauma as well as in the peripartum period.¹⁵⁶

The SIJ is innervated by the posterior rami of primarily S1 and S2 via the lateral branches. The L5 dorsal ramus and the S3 lateral branch have a documented contribution, but to a lesser extent. S4 has been shown to contribute in some studies, but not in others.^72,²⁰¹ The lateral branches were found to innervate the posterior joint capsule and ligaments.

Indications

No historical features or physical examination tests for SIJ pain have been validated with confirmatory SIJ blocks. SIJ-mediated pain also cannot be diagnosed by imaging. Bone scintigraphy has the best specificity of 90% to 100%, but a sensitivity of only 12% to 46%.¹⁷⁷ Consequently diagnostic blocks are the only means of diagnosis.^∗

If the patient has pain above the level of L5 and does not have pain over the sacral sulcus (SIJ) sulcus or point to the posterior superior iliac spine as the primary pain source, then the likelihood of SIJ-mediated pain is low (<10%).⁶⁴ It was noted, however, that SIJ pain can refer into the lower limb. If none of the six provocation tests (distraction, compression; thigh thrust; Gaenslen’s; flexion, abduction, external rotation [FABERs]; sacral thrust) are positive, the SIJ can be ruled out as a pain source.¹¹⁸ If maximal pain is below the level of L5 and three or more of these six provocation tests are positive, the PPV of having SIJ pain is 77%.¹⁸⁸

The utility of an intraarticular SIJ injection with corticosteroids is debatable. As noted below, evidence supports its use in those with inflammatory sacroiliitis. However, in others, the etiology might not be inflammatory, and corticosteroids might be of little or no benefit. Consequently SIJ injection should be performed only after a trial of and in conjunction with noninterventional care.

Technique

Fluoroscopy is used to identify the inferior border of the joint at a point where the anterior and posterior joint lines overlap. The exact angles to find this hyperlucent area vary from patient to patient. Nonfluoroscopically guided SIJ injections have been shown to be intraarticular only 22% of the time, with undesired epidural flow 24% of the time. Consequently fluoroscopy is required to perform an SIJ intraarticular injection.¹⁶² SIJ injections done without fluoroscopy are properly described as ligamentous or trigger point injections.

Once the spinal needle is advanced to the inferior border of the joint and intraarticular placement is expected, needle placement is confirmed with radiopaque contrast and fluoroscopy (Figure 25-8, A). Views are taken in multiple planes to be certain flow is intraarticular and nonvascular. A lateral view (Figure 25-8, B) is taken to show whether there is a ventral defect in the capsule that could allow spread of the medication to the lumbosacral plexus. Extraarticular extravasation has been cited at 61%.⁷³ A corticosteroid and anesthetic mixture is usually injected until capsular resistance is encountered (0.8 to 2.5 mL).⁷³

FIGURE 25-8 Sacroiliac joint intraarticular injection. A, Anteroposterior view. B, Lateral view.

Documentation of quantitative pain relief and postprocedure provocative maneuvers during the anesthetic phase are essential when administering diagnostic injections. Patients then continue to rate their pain at regular intervals in a pain diary, with a positive response demonstrating a duration of relief that is consistent with the anesthetic used.

Evidence

In those with true sacroiliitis, three of four RCTs show benefit from intraarticular corticosteroids.^31,^133,¹⁴³ Evidence of therapeutic efficacy of intraarticualr injection for other causes for SIJ pain is not conclusive, although one study suggests additional extraarticular pain generators such as the posterior interosseous ligaments. This study showed a modest benefit of a combined intraarticular and extraarticular technique.³¹ There is conjecture that the posterior ligaments are a major contributor to SIJ pain.

The ability to reliably block the innervation of the SIJ has not been demonstrated. Controlled studies have shown that the diagnosis of SIJ pain cannot be made via attempted blockade of the lateral branch divisions of the L5–S3 dorsal rami.⁶¹ Intraarticular injection is used as the diagnostic block for the SIJs until a reliable lateral branch block technique can be established.

Zygapophysial and Sacroiliac Joint Summary

• Single intraarticular injections or MBBs cannot reliably diagnose zygapophysial or SIJ pain.

• In the cervical, thoracic, and lumbar spine, using intraarticular blocks instead of MBBs is associated with an increased FP rate and a decreased PPV.

• Relief of at least 80% of the pain is recommended from the first MBB or SIJ injection to proceed with a second confirmatory block; less than 80% relief increases the rate of FP and decreases PPV.

• True comparative blocks enhance specificity (88%), but at the sacrifice of sensitivity (54%) as judged against placebo.

Radiofrequency Neurotomy

Background

Radiofrequency neurotomy (RFN) of the medial branches of the dorsal primary rami aims to destroy the afferent nerve supply to the zygapophysial joints via heat lesioning. It uses radiofrequency current to create heat. The lesion is performed at 85°C for 60 to 90 seconds.¹⁹⁴ Nerve regeneration is assumed to occur in 9 to 12 months, and pain might recur. The same lesion can be repeated, however, and again pain relief is usually obtained.^99,¹⁶⁷

RFN has been used for structures other than the medial branches for zygapophysial joint pain. Other studies have included RFN of the lateral branches of the dorsal primary rami of S1–S4 for SIJ pain and the dorsal root ganglion, which is beyond the scope of this chapter.

Thermal RFN uses a noninsulated electrode tip delivering electrical current to the tissue. Heat is generated as a result of the current flow through the tissues, which causes charged molecules to oscillate rapidly, producing friction. The lesion produced is an ellipse along the length of the noninsulated tip with a diameter 1 to 1.5 times the size of the tip. Accordingly, needle placement parallel to the target nerve is required.¹²⁰

Pulsed Radiofrequency

A potential risk of thermal RFN is that it might cause a painful neuritis from coagulation of the dorsal root ganglion. Some practitioners have proposed that spinal instability results from denervation of the multifidi muscles.¹⁴⁶ Pulsed RF, performed at temperatures not exceeding 42° C, has been proposed as a method to achieve pain interruption with less risk. A study of pulsed RF in the cervical and lumbar spine reported a 60% success rate (pain reduction >50%), with pain relief lasting approximately 6 months.¹⁴⁶ Consequently the routine use of pulsed RFN is not recommended because thermal RFN has shown much better outcomes with little risk.

Technique

Proper patient selection is a prerequisite for RFN as discussed below. Fluoroscopic guidance with the use of radiopaque contrast is required. RFN is often more painful than other spinal procedures, and conscious sedation is typically administered. Physiologic monitoring is mandatory. The needle electrode is positioned over the proper osseous anatomy, parallel with the nerve, using fluoroscopic guidance (Figure 25-9). Motor and sensory stimulation is often used to rule out placement near the nerve root or dorsal root ganglion. The nerve is anesthetized, and finally several overlapping lesions are performed.

FIGURE 25-9 Third occipital nerve radiofrequency neurotomy. A, Anteroposterior view. B, Lateral view.

A cadaveric study of 30 nerve locations in six cadavers determined precise placement for RFN, and it was recommended that needles larger than 22 gauge with a 10-mm active tip would be the most effective.¹²⁰

No techniques for the C0–C1 or C1–C2 articulations have been studied. For C2–T1, the locations of the medial branch nerves are more variable than in the lumbar spine, and the specific location is dependent on the vertebral level. No techniques for thoracic medial branch RFN have been validated, but RFN is performed based on the cervical and lumbar literature.

A postprocedure evaluation is carried out with the patient, who is usually prescribed oral analgesics for postprocedural pain that can last 2 to 3 weeks.

Complications

Serious complications of RFN are rare. Inadvertent placement of the probe near the dorsal root ganglion or nerve root, however, can cause serious injury.¹ Proper needle placement as established by multiplanar imaging, motor stimulation prelesioning, and a responsive patient provide for safe lesioning.

Excluding treatment of the TON, review of 83 cervical RFN procedures documented postoperative pain in 97%, ataxia in 23%, numbness in 80% at C3–C4 and in 19% at lower levels, and dysesthesias in 30% at C3–C4 and in 17% at lower levels.¹²⁸ In 51 RFN procedures to the TON, side effects included numbness in 97%, ataxia in 95%, dysesthesias in 55%, hypersensitivities in 15%, and pruritus in 10%. Most of these side effects were limited to 2 weeks postprocedure and lasted up to 4 weeks in one patient.⁸⁶ Complications of lumbar RFN have been studied, and in 116 procedures, the incidence of local pain lasting more than 2 weeks was 0.5%, and the incidence of neuritic pain lasting more than 2 weeks was 0.5%.¹¹⁵

Conjecture that RFN could cause multifidus atrophy and therefore spinal instability ¹⁴⁶ is unfounded. A recent observational study, although small, did not show identifiable segmental multifidus atrophy by MRI 17 to 24 months after successful lumbar medial branch denervation with RFN despite segmental denervation documented by electromyography (EMG) 6 weeks after RFN. Patients had sustained pain relief.⁶⁶

Evidence

Evidence supporting cervical RFN is substantial. Lord et al.¹³¹ studied 24 patients with chronic neck pain selected from comparative MBBs and lack of relief with saline injections under double-blind conditions. With a 22-gauge needle, RFN with four to six lesions per medial branch was performed to accommodate for variations in nerve position. Eight percent of the control group and 58% of the active group had greater than 90% relief, and the median time until pain returned to a level of at least 50% of the pretreatment level was 263 days (8.8 months) in the active group and 8 days in the placebo group (p = 0.04). Stovner et al.¹⁸⁴ compared sham and active RFN in patients with cervicogenic headache, but did not use any diagnostic blocks for selection. Results did not show statistical significance, highlighting the need for proper patient selection.

A second RFN is successful for patients who have relief from the initial RFN. If patients had at least 90 days of relief from the first RFN, then the chance of a second RFN being successful RF was 82%. If pain relief from the first RF was for less than 90 days, then the success rate for a second RFN was only 33%.¹⁴⁴

Neurotomy of the TON for pain mediated by C2–C3 had poor outcomes initially. However, with a new technique using a larger 16-gauge needle and multiple parallel lesions to account for variability of nerve location, the success rate was doubled from 40% to 80%. Govind et al.⁸⁶ treated 49 patients who had been selected with comparative TON blocks, and achieved complete relief of pain for at least 90 days in 88% of the patients. The median duration of relief was 297 days.

Unfortunately, much of the literature available on the efficacy of RFN has used a single-block technique to identify patients for study inclusion, particularly in the lumbar spine. By using a single MBB technique with only 50% relief required, the best results are greater than 50% pain relief in 46% of the active group versus 12% of the sham group at 1 year.¹⁸⁹ Other studies that have used single intraarticular blocks without high-grade relief for patient selection fare worse, with no active patients having obtained benefit after RFN.^122,¹⁹¹ When substantially more stringent selection criteria are used, however, favorable results have been documented. In a study by Dreyfuss et al.,⁶³ of 41 patients who met the initial inclusion criteria, only 15 were included for lumbar RFN after modified comparative blocks. After RFN the mean visual analog scale (VAS) score dropped 88% (VAS score, 0.5) at 6 weeks and 76% (VAS score, 1) at 1 year. Eighty-seven percent of patients had at least 60% relief at 1-year follow-up and did not seek ancillary treatment for their pain. The small number of patients extracted (15 of 138 total patients with back pain [11%]) for lumbar RFN from the initial population with nonradicular low back pain reflects the low prevalence of pure zygapophysial joint pain in the lumbar spine. A double-blind RCT was performed by Nath et al.¹⁴⁹ comparing lumbar RFN with sham treatment in patients with modified comparative block–proven zygapophysial-mediated pain, establishing that lumbar RFN is superior to placebo.

Appropriate RFN technique is also relevant to its efficacy. Lesion size sufficient to denervate the medial branch within its osseous territory is far more likely by using larger radiofrequency probes (16 to 18 gauge) and with a longer (e.g.,10 mm) active tip. The lumbar RF study by Dreyfuss et al.⁶³ used a 16-gauge, 5-mm tip needle, but two contiguous lesions were performed to ensure a 10-mm-long lesion. Technical success, as measured by presegmental- and postsegmental-specific multifidus EMG, was 90%. The study by Nath et al.¹⁴⁹ used a 22-gauge electrode needle with a 5-mm active tip, and two contiguous lesions (equating to a 10-mm lesion) were made at three different locations for each medial branch. The inaccuracies of the RFN technique used in other studies could have had a negative effect on the outcomes.

Sacroiliac Joint Radiofrequency Neurotomy

Background

Because intraarticular sacroiliac joint (SIJ) injections have not offered long-term relief of SIJ-mediated pain, RFN has been postulated for longer term pain management because of its effectiveness in the cervical and lumbar spine. Retrospective studies of CT or fluoroscopically guided “denervation” along the posterior SIJ line or in the interosseous ligament did not show benefit. This technique is flawed, however, because the lateral branches are not accessible at these locations.^70,⁸⁴

Technique

Overall the technique is similar to that for medial branch RFN procedures. The lateral branch anatomy, however, is less consistent. New techniques are currently being developed to obtain more reliable lesioning, which is beyond the scope of this chapter.

Evidence

Cohen and Abdi ⁵² reported retrospectively in 2003 on 18 subjects who had greater than 80% relief with SIJ intraarticular blocks. Nine of the 18 subjects had greater than 50% relief from L4 medial branch, L5 dorsal ramus, and S1–S3 lateral branch blocks. Eight of these 9 patients who subsequently underwent SIJ RFN had greater than 50% pain relief that was persistent at 9 months’ follow-up. A prospective cohort study of bipolar lesioning showed little result for decreasing pain.⁴¹ Consequently the evidence for RFN to manage SIJ pain is still quite limited.

Summary

• Optimal RFN results occur when very stringent selection criteria are used.

• Technical success is maximized with parallel needle placement, larger gauge needles, and a multiple lesion technique to account for variable nerve location.

• The PPV for SIJ injections is much less and the occurrence of FP responses much higher than for MBBs. The use of an anatomic control injection (e.g., MBBs with or without discography) or a placebo injection should be strongly considered before SIJ RFN.

Discography

The differential diagnosis of axial low back pain is presented in Chapter 40. Interventional strategies are used for both the diagnosis and management of diskogenic pain. Patients with degenerative disk disease diagnosed via imaging, such as MRI or CT, can be asymptomatic.^23,¹⁰² No physical examination finding is diagnostic for diskogenic pain, and it is difficult to attribute a specific disk level to these symptoms. Abnormal-appearing disks have also been seen in both symptomatic and asymptomatic subjects, leading to uncertainty about whether a specific disk can be precisely pinpointed as a focal pain generator.^11,^76,⁸⁷ Lumbar provocation discography is a technique designed to attempt to diagnose a “symptomatic” disk.

Discography was originally described by Lindblom to diagnose a herniated lumbar intervertebral disk, and has been used clinically since the 1940s. During early discography procedures, patients were noted to commonly report reproduction of their typical pain ²⁵ with stimulation, and abnormal disks were rarely painful in asymptomatic controls.¹⁴²

Lumbar discography is known by several names, including lumbar disk stimulation or provocation discography. Because this procedure serves as both an imaging and provocation study, provocation discography should be considered the valid term. The ultimate goal of lumbar provocation discography is to identify a painful lumbar intervertebral disk. It is purely a diagnostic study. Discography involves insertion of a needle into the nucleus pulposus of the target disk, followed by the instillation of contrast dye to outline the morphology and evaluate for internal disruption. Disk provocation refers to the recording of a patient’s response to the injection. Manometry is typically incorporated to measure opening pressure and intraprocedure pressure changes. Additional information regarding the morphology of the disk can be obtained with postinjection CT.

The premise of the procedure is that if a disk is indeed symptomatic, provocation should reproduce the usual, or concordant, pain. If stressing that disk either does not cause pain, or produces pain that is not concordant with the usual pain, then the etiology of the patient’s pain cannot be attributed to that disk. However, even if stimulation of the disk results in concordant pain, it does not reveal the pathoanatomic nature of that pain.

According to the International Association for the Study of Pain and the International Spine Intervention Society, diagnostic criteria for diskogenic pain using provocation discography are as follows: (1) provocation of the target disk reproduces the patient’s typical pain, and (2) provocation of adjacent disks does not reproduce pain. Pain provocation at control levels yields an equivocal test, as does discordant pain. As with other provocation tests, there can be an inherent issue with FP results. In an attempt to minimize this concern, intervertebral disks at adjacent levels should be tested, to serve as controls. Pain at adjacent levels does not necessarily exclude diskogenic pain, but the diagnosis cannot be made with certainty.

Risks, contraindications, and potential complications are similar to those common to all interventional spine procedures. Unique to discography is a higher risk for infection.

Antibiotic prophylaxis was not routinely used for lumbar discography in the past. At that time the rate of diskitis was reported to be 0.1% to 2.3% per patient, and 0.05% to 1.3% per disk.²⁶ Typically infections were caused by Staphylococcus aureus, Staphylococcus epidermidis, and Escherichia coli. With current techniques, including routine prophylactic antibiotics, the rate of diskitis has been lowered.⁷⁵

The validity of discography has been evaluated in several studies, but still remains controversial. Discography has been performed in both symptomatic and asymptomatic subjects, to analyze its validity and concordance. Walsh et al.¹⁹⁷ in 1990 reported a FP rate of 0% in asymptomatic controls, and a true-positive rate of 89% in symptomatic subjects. Conversely, Carragee et al.⁴⁷ in 2006 reported that the best-case PPV for discography was 50% to 60%. More recently Wolfer et al.,²⁰⁴ in a metaanalysis of 11 studies, reported FP rates of 9.3% per patient and 6.0% per disk. In data pooled from asymptomatic patients, the FP rate was 3.0% per patient and 2.1% per disk. They reported an improvement in the FP rate with low-pressure positive criteria (≤15 psi). They calculated an overall specificity of 94% (95% confidence interval, 0.88 to 0.98) and a FP rate of 6%. Another study has suggested discography as a potential cause of accelerated disk degeneration changes.⁴⁵

The usage of discography remains controversial, but in properly selected patients with proper technique and standards it appears to have diagnostic value.

Summary

• The goal of lumbar provocation discography is to identify a painful lumbar intervertebral disk.

• Abnormal-appearing disks are seen in both symptomatic and asymptomatic subjects.

• Risks, contraindications, and potential complications are similar to those common to all interventional spine procedures, but also includes a higher risk for infection.

• The validity of discography remains controversial.

Intrathecal Delivery Systems

Intrathecal drug delivery systems such as implantable subcutaneous pumps are being used for patients who have chronic pain that is resistant to management with oral or parenteral medications. As early as 1979, both epidural and intradural opioid delivery systems had been reported in the literature.^17,¹⁹⁹ Delivery of agents directly to opioid receptors was noted to decrease the systemic effects of the medications and also to decrease the dosage needed to achieve analgesia.

For pain treatment, three classes of medications are used for intrathecal delivery: opioids, local anesthetics, and nonopioid agents such as adrenergic receptor agonists or NMDA (N-methyl-D-aspartate) receptor antagonists. Infusion of opioids is the most frequently used.¹⁵⁷ Intrathecal morphine infusion has been shown to be safe at moderate concentrations, with favorable efficacy data.^18,^150,¹⁵³ There is, however, increasing interest in using hydromorphone because it is five times as potent as morphine with a better side effect profile.¹⁵⁰ Bupivacaine and clonidine also demonstrate favorable efficacy and toxicology.^90,⁹¹

According to Prager,¹⁵⁵ indications for intraspinal drug delivery include chronic pain with known pathophysiology, sensitivity of pain to the medication being used, failure of more conservative therapy, a favorable psychosocial evaluation, and a favorable response to a screening trial. Contraindications include systemic infection, coagulopathy, allergy to the medication being used, inappropriate drug habituation (untreated), failure to obtain pain relief in a screening trial, unusual observed behavior during the screening trial, poor personal hygiene, and poor patient compliance. Although rare, granuloma formation is a potentially severe complication of intrathecal delivery systems.

Summary

• Intrathecal drug delivery systems are used for patients with chronic pain that is resistant to management with oral or parenteral medications.

• Opioids, local anesthetics, and nonopioid agents (i.e., adrenergic receptor agonists or NMDA receptor antagonists) can be used for intrathecal delivery.

Spinal Cord Stimulation

Spinal cord stimulation is a procedure using electrodes placed within the epidural space adjacent to the area of the spinal cord or root presumed to be the source of pain. An electric current is then applied to achieve sympatholytic and other neuromodulatory effects.⁵⁰ The premise of spinal cord stimulation is based on the gate control theory of pain.⁴² The “gates” that control nociceptive input can be “closed” via stimulation of low-threshold primary afferent fibers.

Studies evaluating spinal cord stimulation in patients who have failed back syndrome with persistent radicular pain have demonstrated positive outcomes, including decreased opioid use and increased pain relief as compared with reoperation.^117,¹⁵² Randomized trials for the treatment of low back pain using spinal cord stimulators in nonfailed back syndrome, or failed back surgery syndrome without radicular pain, are not found in the current literature.⁵⁰

Summary

• Based on the gate control theory of pain, spinal cord stimulation uses an electric current for pain management.

• Spinal cord stimulators might have a role in persistent radicular pain after surgery.

Sympathetic Blocks

The study of the sympathetic nervous system began with Galen in the second century AD. Much of the work on sympathetic nerve blocks blossomed from necessity of treatment of traumatic casualties during World War II. In 1953 Bonica ¹⁰³ proposed guidelines for neural blockade.

To understand the proper usage of sympathetic blocks, one must study the basic anatomy of the sympathetic nervous system. Briefly, the autonomic nervous system is composed of the sympathetic and parasympathetic systems. Cervical ganglia are located ventral to the transverse processes; thoracic ganglia reside ventral to the head/neck of the ribs; lumbar ganglia are located anterolateral to the vertebral bodies and anteromedial to the psoas muscle; and sacral ganglia lie on the anterior sacrum, medial to the anterior sacral foramina.

Sympathetic blocks interrupt both efferent (sympathetic motor) and afferent fibers, causing interruption of nociceptive pathways as well as the sympathetic vasomotor, sudomotor, and visceromotor nerves. Low concentrations of local anesthetic block small myelinated Aδ fibers, unmyelinated C fibers, and preganglionic efferent sympathetic B fibers, but do not block somatic motor nerves.

Nerve blocks use two different medications: anesthetic agents and neurolytic agents. Local anesthetics work through a reversible block of the axon’s sodium channels. The extent of blockade depends on the specific characteristics of the medication used. Neurolytic agents used in the United States include alcohol and phenol, which block both the sensory and motor nerves. Because of its onset of action, a 12- to 24-hour observation period is required before determining the efficacy of an alcohol block. Patients often report an intense burning on injection. Radiographic contrast can also be added to view the spread of the injectate. Phenol does not produce the same burning side effects on injection as alcohol does. Because it has local anesthetic effects, it can result in pain relief immediately on injection; but long-term pain relief cannot be determined for 24 to 48 hours postinjection.

Stellate ganglion blockade can be used for both diagnostic and therapeutic purposes for sympathetic-mediated pain. Typical indications include complex regional pain syndrome (CRPS) I and II, acute herpes zoster, and phantom limb pain causing symptoms involving the head, neck, face, and upper limbs. (It should be noted that sympathetic block in CRPS is controversial because of the lack of evidence that the sympathetic system is involved in CRPS. This is one of the reasons the International Association for the Study of Pain changed the name of reflex sympathetic dystrophy to CRPS I.) The stellate ganglion is located anterior to the transverse process of C7, posterior to the neck of the first rib and just below the subclavian artery. Specific techniques are beyond the scope of this chapter. However, given the proximity of the stellate ganglion to vital structures, fluoroscopy should be used, and patients should be fully monitored during the procedure.

Lumbar sympathetic blocks can be used both diagnostically and therapeutically for pain in the pelvis and lower limbs related to CRPS and phantom limb pain. The lumbar sympathetic ganglion is located along the anterolateral surface of the lumbar vertebral bodies, anteromedial to the psoas muscle. Injection at L2 blocks sympathetics to the bilateral lower limbs and the pelvis. These procedures should be performed with the assistance of fluoroscopy, and the patients should be monitored throughout.

The ganglion impar, also called the ganglion of Walther, is located in the sacrococcygeal region. The pelvic portion of each sympathetic trunk runs anterior to the sacrum, medial to the anterior sacral foramina. The two pelvic sympathetic trunks converge at the distal sacrum and terminate on the anterior surface of the coccyx in a small ganglion called the ganglion impar. Foye et al.⁷⁴ reported that a single local nerve block of the ganglion of impar can provide 100% relief of coccydynia.

As with any interventional procedure, there is a risk of complications. Accidental subarachnoid or epidural blockade can occur with any injection performed close to the spine, and injections in the cervical or thoracic region carry a risk of pneumothorax from an improperly placed needle. In the lumbar spine, large vessels can be inadvertently punctured.

Summary

• Sympathetic nerve blocks use anesthetic agents and/or neurolytic agents for diagnostic, prognostic, and/or therapeutic purposes (Table 25-4).

Table 25-4 Summary of Sympathetic Nerve Blocks

Function	Indication
Diagnostic	1. Ascertain specific nociceptive pathway 2. Define mechanism for chronic pain 3. Elucidate differential diagnosis of pain 4. Distinguish between true sympathetic pain and partial sympathetically mediated pain