Industrial Medicine and Acute Musculoskeletal Rehabilitation. 5. Interventional Procedures for Work-Related Lumbar Spine Conditions

5.1 Clinical Activity: To discuss the differential diagnosis and procedural management of a 40-year-old dockworker with low back pain and leg pain for whom conservative treatment has failed 

TO ESTABLISH AN APPROPRIATE treatment plan for an injured worker with low back pain (LBP) and referred leg pain, it is imperative to develop a proper differential diagnosis. Clearly, this process is guided by the history, physical examination and laboratory and imaging studies. The clinician must evaluate and rule out systemic diseases including cancer and rheumatologic disease, which may have been exacerbated by activity at work or home. Obtaining a detailed medical history, review of systems (including constitutional symptoms), and psychosocial background are critical in determining an appropriate differential diagnosis. After systemic diseases have been ruled out, the differential diagnosis then dictates the treatment algorithm and appropriate diagnostic and therapeutic interventions.

In this patient with both LBP and referred leg symptoms, the differential diagnosis should include but not be limited to lumbar disk herniation, sacroiliac joint pain, zygapophyseal joint pain, diskogenic pain, piriformis syndrome, spondylolisthesis, lumbosacral plexopathy, and in today’s aging work force, exacerbation of lumbar spinal stenosis. To determine which spinal intervention is most appropriate in these cases, one must consider the goal of the intervention. Some spinal interventions can be useful diagnostically, whereas others are therapeutic. For persistent pain presumed to emanate from the zygapophyseal joint or the sacroiliac joint, intra-articular injections are appropriate treatment options.

The prevalence of zygapophyseal joint−mediated pain, commonly referred to as a facet syndrome, varies between 15% and 40% of chronic LBP patients.1, 2 Intra-articular zygapophyseal-joint injections can be used therapeutically and may also be diagnostic if pain relief is achieved. For a more purely diagnostic approach, performing anesthetic blocks of the medial branches of the dorsal rami supplying the zygapophyseal joint in question may help ascertain the exact level of involvement and may also help correlate the patient’s history, physical examination, and imaging study results. In the double-block approach, injections targeting the medial branches of the dorsal rami, commonly referred to as medial branch blocks, are performed at 2 separate occasions. Different anesthetics with different durations of action are administered at the 2 occasions. Pain relief lasting at least the minimum length of effectiveness of the local anesthetic constitutes a positive response. This technique is considered superior to single anesthetic blocks in the diagnosis of zygapophyseal joint−mediated pain and is the criterion standard for making the diagnosis. Based on a beneficial response to these injections, one might consider radiofrequency ablation of the appropriate medial branches.

The prevalence of sacroiliac joint pain ranges from 18% to 30% of chronic LBP patients. Intra-articular sacroiliac joint corticosteroid injections remain the only interventional option proven useful in this condition.3, 4, 5 However, these injections have been controversial. Diagnostic injections using the comparative local anesthetic block approach may also help confirm the role of the sacroiliac joint as a pain generator.

There are 3 approaches for performing epidural injections: caudal, interlaminar, and transforaminal. Lumbar spinal stenosis, lumbar disk herniation, and lumbar diskogenic pain may all present as LBP with or without a radicular component and are common diagnoses for which an epidural steroid injection (ESI) is routinely performed. To date, prospective, blinded, and controlled scientific studies supporting the efficacy of lumbar ESIs for the treatment of these disorders have been lacking. Certain patient populations warrant specific mention. For patients with lumbar stenosis, interlaminar ESI should be performed below the level of stenosis to minimize the risk of intrathecal injection. Interlaminar ESI may not be helpful in the treatment of neuroforaminal stenosis, whereas a transforaminal ESI may be more appropriate.

5.2 Educational Activity: To discuss the biomechanic changes and treatment recommendations for this 40-year-old dockworker diagnosed with a lumbar disk herniation 

Various treatment options exist for lumbar herniated nucleus pulposus (HNP), including medications, therapies, injections, surgical diskectomy, and fusion. Although there have been multiple treatments proposed for a lumbar HNP, the natural history is favorable for spontaneous improvement. In a controlled, matched cohort of patients with lumbar HNP, the group treated conservatively had similar pain improvement at the 10-year follow-up. Further evaluation showed that the surgical group had a slightly better result for complete relief of pain than did the conservatively treated patients.6 No specific treatment has had a clear advantage over another. After an appropriate trial of more conservative treatments including medications, physical therapy, and activity modifications (but not bed rest), one may consider interventional procedures if pain persists. To understand the rationale behind these treatments, one should have a basic understanding of the biochemistry and pathophysiology associated with a lumbar HNP.

Radicular pain as a result of lumbar HNP has been described.6 Although initial studies focused on mass effect and pressure on the nerve root created by HNP, more recent research7, 8, 9, 10, 11, 12, 13 has focused on the subsequent inflammatory milieu. Studies7, 8, 9, 10, 11, 12, 13, 14 showed that neural compression can cause painless weakness, and many studies7, 8, 9, 10, 11, 12, 13 have confirmed that the nucleus pulposus contains a multitude of inflammatory mediators. The combination of mechanical compression of the nerve root and presence of inflammatory mediators seems to be more than just additive.7 Researchers have searched for the inflammatory markers associated with a lumbar HNP. Phospholipase A2 was found within the extravasated nucleus pulposus,8 validating the theory that nuclear material can cause inflammation. Several other inflammatory markers have been associated with the nucleus pulposus including nitric oxide, interleukin-6, prostaglandin E2, and matrix metalloproteinases. These markers may also be tied to increases in interleukin-1.9, 10 The isolation of tumor necrosis factor alpha (TNF-α) has been inconsistent in the setting of disk herniation. The presence of both leukotriene B4 and thromboxane B2 has also been associated with the HNP.11

In various animal studies of lumbar HNP, changes in nerve function and structure, cellular matrix, and inflammatory markers have been observed.7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17 These studies examined both experimentally induced disk disruption in vivo and exposure to nucleus pulposus materials in tissue culture. Leukotaxis and increased vascular permeability have been noted in vitro as a result of exposure to nucleus pulposus material.12 Different researchers have noted an increase in macrophages and a possible increase in mast cells in the inflammatory milieu.13 Of note, in nerve tissue surrounded by nucleus pulposus, myelin changes occur in the local area of exposure. Other changes include an increase in Schwann cell cytoplasmic and intracellular edema.15 Decreases in nerve conduction velocity (NCV) and amplitude of the compound muscle action potential are also seen.7 The presence of TNF-α has been correlated with a decrease in NCV in nerve roots that have been exposed to nucleus pulposus. One animal study16 has shown that application of the nucleus pulposus decreases the blood flow to the dorsal root ganglion, with increased edema of the dorsal root ganglion and interneural space. This may be a result of increased vascular permeability mentioned earlier. Other researchers have found a direct relation between blood flow to the nerve and NCV. There is a time lag between the exposure and the measured decrease in NCV.17

In accordance with the described pathophysiology, corticosteroids and anesthetics have been used for injection into the epidural space in the setting of lumbar HNP. Studies have shown that corticosteroids may have more than anti-inflammatory effects. Corticosteroids may have some anesthetic affects on small unmyelinated C fibers in irritated neural tissue.18 In porcine models, high doses of methylprednisolone given early after an exposure to nucleus pulposus material showed a protective effect on NCV in the exposed nerves.19 Lidocaine has also shown a mild protective effect in nerves exposed to nucleus pulposus when evaluating the NCV. It seems to have its greatest effect when administered early. Studies have also suggested that lidocaine may exhibit anti-inflammatory properties such as decreasing phagocytosis, polymorpholeucocyte lysosomal enzyme release, and superoxide anion production.18, 20, 21

ESIs are used to deliver anti-inflammatory medications and local anesthetics in concentrated amounts to the epidural space surrounding areas of presumed pathology and inflammation. Caudal and interlaminar approaches were the initial routes of administration of medication in epidural injections pioneered before the advent of fluoroscopy. The advantage of these approaches is easy access to the epidural space. Without fluoroscopy, however, no approach allowed confirmation of successful deposition of medication into the epidural space. Fluoroscopic studies of interlaminar injections performed without imaging (blind) have shown that, even in experienced hands, nearly one third of blind injections do not truly access the epidural space.22 The efficacy of blind interlaminar injections has subsequently been questioned. Lack of fluoroscopic confirmation of epidural flow on injection of contrast medium also likely increases the risk of dural puncture. Fluoroscopically guided ESI with contrast confirmation improves the accuracy and—by supposition—the efficacy of these injections. Retrospective and limited prospective trials have shown generally short-term pain relief in heterogeneous patient populations with questionable long-term benefits.23 The lack of well-controlled, prospective trials assessing the efficacy of interlaminar ESI, with or without fluoroscopic guidance, leaves the true efficacy of these injections open to debate. Further, this literature rarely includes functional outcome measures beyond pain relief.

Compared with interlaminar ESI, transforaminal ESI has a theoretic advantage of delivering medication to the anterior epidural space where the nerve root traverses. By necessity, these injections are performed with fluoroscopic guidance and, ideally, with contrast medium to confirm needle placement within the epidural space and to prevent injection inadvertently into vascular structures. Several studies24, 25, 26 evaluated the effectiveness of these injections for a lumbar HNP. In particular, recent prospective, controlled studies show that transforaminal ESI may help patients avoid surgery and reduce pain in the setting of lumbar HNP.24, 25

Given these studies and the evolving evidence regarding the biomolecular effects and pathophysiology of lumbar HNP, it is reasonable to recommend ESIs performed with fluoroscopic guidance and contrast confirmation as appropriate treatments for HNP. The literature suggests potential advantages with a transforaminal approach; however, currently no well-controlled comparative study evaluates the various epidural approaches for delivering medications.

5.3 Educational Activity: To describe, in a 35-year-old home improvement warehouse worker with predominantly LBP, the use of interventional procedures in the diagnosis and treatment of posterior element pain 

The posterior elements of the lumbar spine include the zygapophyseal joints and the sacroiliac joints. The use of interventional techniques to confirm the diagnosis of zygapophyseal joint− or sacroiliac joint−mediated pain and interventional treatment options for these conditions are outlined as follows.

Revel et al27 described a set of criteria increasing the likelihood of a diagnosis of zygapophyseal-joint pain in the lumbar spine, including age greater than 65 years; pain well relieved by recumbency; and no pain exacerbation with coughing, sneezing, forward flexion, lumbar extension, rising from a flexed lumbar spine, or extension combined with rotation. Other investigators28 have challenged these criteria, claiming that these screening tests yield a low sensitivity and a high specificity for the diagnosis of zygapophyseal-joint pain. A single physical examination maneuver or combination of maneuvers correctly identifying the zygapophyseal joints as the primary source of pain remains to be elucidated.1, 29 Other methods are often used to confirm the clinical suspicion of zygapophyseal-joint pain. The zygapophyseal-joint pain referral patterns, based on provocative injections, may include the lower back and ipsilateral hip.2 Clinicians often use controlled comparative local anesthetic blocks. These are a more objective way to confirm the diagnosis of zygapophyseal-joint pain. The dual innervation of the lumbar zygapophyseal joints (fig 1) must be considered when contemplating medial branch blocks.

View Large Image Download to PowerPoint

Fig 1. Left posterior view of the midlumbar spine showing the dual innervation of the zygapophyseal joints by paired articular branches from the medial branches of the dorsal rami. Legend: a, articular branch; DR, dorsal ramus; ib, intermediate branch; mb, medial branch; tb, terminal branches; VR, ventral ramus.

Although generally considered diagnostic, medial branch blocks may also be therapeutic, with long-term pain relief achieved in some patients. Therapeutic benefit may result from disruption of the central feedback loop that forms in chronic pain states or by inadvertent neurotomy of the medial branch by the needle tip. Adding steroid to the local anesthetic in these injections may potentiate the anesthetic blockade effects, but no literature supports its use.

The interventional treatment options commonly used to alleviate zygapophyseal-joint pain include intra-articular injections and radiofrequency ablation of the medial branches. Intra-articular zygapophyseal-joint injections with corticosteroids and local anesthetic have been validated as an effective treatment for zygapophyseal-joint pain.2 Drawbacks of zygapophyseal-joint intra-articular corticosteroid injections include limited long-term efficacy. Because of potential deleterious effects of steroids on the synovial lining and ligamentous supports for these joints, as well as potential local osteoporotic effects, clinicians should limit the number of injections a patient receives.

Several studies have confirmed the efficacy and long-term pain relief of radiofrequency ablation of neural structures. In this procedure, the nerve and tissues surrounding the electrode are heated to 80°C, thus causing destruction of the medial branch of the dorsal ramus itself. When it is administered after diagnostic, comparative, local-anesthetic, medial branch blocks, this therapy relieves pain in up to 85% of these patients, providing long-term (median, 263d) relief sometimes lasting longer than 1 year. Workers’ compensation patient outcomes for this therapy rival those of the general population.22 The nerve typically regenerates in 90 days. If the pain returns, a repeat of the radiofrequency ablation generally affords the same long-term relief obtained by the initial treatment. Although the medial branch also innervates a small portion of the lumbar multifidus muscle, no functional weakness has been associated with this procedure. One study30 validated good relief of pain with radiofrequency neuroablation after a single anesthetic block; however, the efficacy is diminished compared with the use of diagnostic, controlled, comparative, local anesthetic medical branch blocks.31

The sacroiliac joint is a true diarthrodial joint containing approximately 2 to 3mL of synovial fluid. Joint pain may result from trauma, cumulative sheer forces (eg, lifting, running, altered gait mechanics), a rheumatologic process such as ankylosing spondylitis or pregnancy, or it may be idiopathic.32 Physical examination maneuvers, particularly in isolation, have limited diagnostic utility. Use of combination maneuvers may increase diagnostic yield.33 Pain maps of the sacroiliac joint show significant overlap with those generated for the lower-lumbar zygapophyseal joints and pain patterns from radiculopathies or diskogenic pain but do not typically include pain above the L5 level. One study suggests pain may refer distally to the feet.34 The double-block procedure has become the diagnostic standard.

Intra-articular injection of local anesthetic and corticosteroid under fluoroscopic guidance is a common interventional treatment for sacroiliac joint pain. Given the technical difficulty of the injection even under fluoroscopic guidance, successful injection of the sacroiliac joint occurs rarely when performed blindly. Injections for sacroiliitis due to ankylosing spondylitis generally last 6 months or less. Patients who had a single anesthetic block that provided pain relief responded well to intra-articular corticosteroid injections.5 These injections afforded long-term relief of up to 12 months, with patients frequently requiring multiple injections (average, 2.1) to achieve results. Intra-articular injections of the sacroiliac joint are best used as part of an integrated rehabilitation program. Newer studies have applied the zygapophyseal-joint double block procedure to select patients for radiofrequency neuroablation to denervate the sacroiliac joint. This treatment may only address the extra-articular pain sensations and not the intra-articular pain associated with sacroiliac joint disorders. A complex and variable innervation of the sacroiliac joint exists compared with that of the zygapophyseal joints. Given this variability and the limited anatomic landmarks for the nerves innervating the sacroiliac joint, consistent denervation of the joint is difficult to achieve. Although radiofrequency neuroablation to treat sacroiliac joint pain had favorable results in a recent, small cohort study,35 further research is needed.

5.4 Educational Activity: To discuss further diagnostic investigation and treatment for the above 40-year-old dockworker with disk degeneration, LBP, and referred leg pain for whom conservative management has failed 

The diagnosis of true diskogenic pain emanating from the intervertebral disk is difficult to definitively establish. In patients with chronic LBP, the prevalence of diskogenic pain is estimated at 40%.36 In an attempt to establish the diagnosis of diskogenic pain, researchers have evaluated several tests. A high-intensity zone on magnetic resonance imaging within the intervertebral disk is presumed to reflect a potentially painful annular tear. These tears may be partial-thickness or complete tears without frank herniation. Because of morphologic changes, increased nociception may occur in a disk with an annular tear. Also, leakage of inflammatory mediators present in the nucleus pulposus may increase nociception. This phenomenon may result from an ingrowth of sinuvertebral nerve fibers as the body attempts to heal the tear. However, the high-intensity zone exists in both symptomatic and asymptomatic subjects, thus clouding its role in the diagnosis of diskogenic pain.

Diskography has been promoted as the de facto criterion standard in the evaluation of diskogenic pain in the absence of a true standard. Diskography is typically performed at 3 levels in the lumbar spine to provide at least 1 theoretic control, or nonpainful, disk level. The procedure is painful, and risks include nerve injury, diskitis, and epidural abscess. Its routine use should be avoided. However, it may have a role in patients with chronic back and/or referred leg pain to spare nonpainful adjacent disks from an intervention, to narrow or establish treatment options, and to serve as yet another correlate to imaging studies. Diskography has limited value for predicting success with surgical fusion.37 Several researchers have noted abnormal results with diskography in asymptomatic patients.38, 39, 40 Other chronic pain states such as fibromyalgia may affect diskography interpretation.41

New interventions, including intradiskal electrothermal annuloplasty, nucleoplasty, percutaneous disk decompression, intradiskal restorative injections, and disk replacement have fueled the diskography debate. Because the validity of diskography in diagnosing true diskogenic pain has been called into question, researchers have focused on isolating variables that may create false positive studies. A recent study with manometric pressure measurements suggests that asymptomatic patients with painful diskograms report only a lower intensity of pain occurring only at very high pressures.40 More recent studies suggest that painful disks at low pressures and with smaller injectate volumes on diskography more likely identify “true diskogenic” pain generators.40 Pressure manometry may also improve interexaminer reliability.

Intradiskal electrothermal annuloplasty was developed based on the theory that annular tear is a cause for diskogenic pain. It has potential advantages over fusion in that it is less invasive, less costly, and it preserves motion of the lumbar segment.

There are 3 proposed mechanisms by which this electrothermal approach may reduce pain and treat the annular tear. Thermal injury applied to the annulus via the heating element catheter may produce collagen remodeling with potential fibrous contraction of the tear. Some researchers question whether the target temperature of 90°C would result in collagen remodeling. A second proposed mechanism is thermal ablation of nerve endings in the outer third of the annulus fibrosus. The disk receives its innervation from the sinuvertebral nerves, which contain A-delta and C fibers. These nerve fibers have been shown to penetrate into the annulus of degenerative disks. This suggests that thermal ablation of these nerves should achieve immediate relief of pain emanating from the disk itself. Although some patients experience immediate relief after intradiskal electrothermal annuloplasty, it may take up to several weeks or longer for the full effects to be realized. The third mechanism is the proposed shrinkage of the nucleus pulposus by the thermal energy.

The indications for intradiskal electrothermal annuloplasty include axial pain with sitting intolerance, lasting longer than 6 months with or without referred lower-limb pain, 50% preservation of the height of the disk when compared with normal disks, and an absence of significant spinal stenosis. Provocation diskography with follow-up computed tomography before treatment may further define the anatomy and may confirm the presence of an annular tear without significant disk protrusion. The symptomatic disk, when provoked, should produce concordant pain at low pressures with control disks exhibiting discordant or no pain.

Contraindications for intradiskal electrothermal annuloplasty may include bleeding diatheses, previous surgery at the potential site of intervention, disk protrusion beyond the posterior longitudinal ligament in proximity to the nerve root origins, and stenosis at the level of interest. Smoking and obesity represent risk factors for a poor outcome after this therapy.

Complication rates are estimated at between 0.7% and 16%36 and are generally self-limited. There have been several reports of avascular necrosis of the vertebral body after intradiskal electrothermal annuloplasty. The mechanism of this injury is unknown. A report42 of a frank disk herniation also exist, as does a report43 of cauda equina syndrome after it. The cauda equina syndrome was attributed to inadvertent catheter placement and heating in the spinal canal. Although infection and thermal injury of the nerve root origins are theoretic risks, there have been no cases reported to date.

In prospective case series of intradiskal electrothermal annuloplasty for chronic LBP, 50% to 81% of patients had good or excellent results. The variable outcomes may result from the use of different outcome scales and inconsistent definitions of “good” or “excellent” results.42, 43, 44, 45 Versus a sham treatment, the electrothermally treated patients in a randomized placebo controlled trial had a statistically significant decrease in pain, although 50% of the annuloplasty group had no relief.45 Based on the Oswestry Disability Index, only 13.5% of the treatment group received more than 75% pain relief. Further analysis suggests greater benefit from intradiskal electrothermal annuloplasty for patients with either very high pain scores or lower levels of function before treatment. Patients with more moderate or mild symptoms benefit less.46 Results in all favorable studies suggest better outcomes with single as opposed to multiple disk levels of treatment. A subsequent study reported no benefit from intradiskal electrothermal annuloplasty when compared with placebo.47

Percutaneous nucleoplasty uses radiofrequency energy to break down molecular structures in a nucleus pulposus. Cadaver studies suggest that nucleoplasty may decrease disk pressures and allow the disk to shrink, thus relieving intradiskal and intraspinal pressure.48

Indications for nucleoplasty may include axial back pain with or without referred leg pain, 50% preservation of disk height, and imaging confirmation of disk protrusion. Contraindications include infection and bleeding diathesis. This procedure is approved for no more than 2 disk levels per treatment. Only a few prospective cohort studies have investigated nucleoplasty. These studies have shown that up to 80% of the patients undergoing nucleoplasty have good to excellent relief of their symptoms.49, 50 Another prospective study48 evaluated patients who underwent nucleoplasty with or without concurrent intradiskal electrothermal annuloplasty. The patients were included if they had referred leg pain with or without back pain. Only 1 in 16 patients had greater than 50% relief. Some patients who had not been using opioid pain medications began such treatment after nucleoplasty. The researchers suggested that different patient selection criteria might have explained the poor outcome in this study. There is a paucity of literature to support nucleoplasty, with numerous critics questioning whether denaturation of central nucleus pulposus material affects in vivo disk pressure or the inflammatory milieu in the outer annulus and spinal canal.

Intradiskal injections of steroids were performed previously and recently reexamined, but no studies to date have shown a positive treatment effect.51 Intradiskal injection of a mixture of glucosamine, chondroitin sulfate, hypertonic dextrose, and dismethylsulfoxide resulted in improved pain based on the Roland-Morris Disability Questionnaire and visual analog pain rating scales.52

Limited study of percutaneous lumbar disk decompression suggests that patients for whom conservative management has failed, have 50% disk height preserved, and experience radicular pain associated with a contained herniation less than or equal to 6mm may benefit. Disk material is removed through the introducer needle after a drill-like bit is threaded through the needle, removing up to 1 to 2mL of nuclear material. The best study to date showed a 65% decrease in pain.53 However, this procedure is relatively new and needs further investigation.

Based on research54 that suggested a role for TNF-α in the pathogenesis of pain from an HNP, intravenous infliximab (a monoclonal antibody inhibiting TNF-α) has been used to treat HNP. An initial open label study55 of a small number of patients with sciatica from HNP had promising results; however, the results of a randomized controlled study56 did not corroborate the initial study findings.

Disk arthroplasty (complete replacement of the disk) was first approved in the United States in 2004. The first-ever disk replacement was performed in Europe in 1984. Disk arthroplasty theoretically preserves physiologic and functional motion of the spinal segment. Since its inception, over 100 prosthetic disk types have been designed and manufactured; however, only 10 have been implanted into human subjects to date.57 In general, disk replacements are composed of a sliding polyethylene core affixed to the vertebra bodies by 2 metal implants. Some designs have a metallic core. Outcomes of initial studies suggest that lumbar disk arthroplasty was at least as efficacious as lumbar fusion. Approximately 80% of the patients reported excellent results in these studies. Some have correlated this favorable clinical outcome with preservation of motion of the spinal segment.58

The clinical indications for lumbar disk arthroplasty are fairly broad and include back pain, with or without leg pain, isolated to no more than 2 symptomatic disks. Provocation diskography may help determine the appropriate disk levels for treatment but has the aforementioned diagnostic limitations. Several factors predict possible failure of this treatment, including osteoporosis, structural deformities, and pain localized to the zygapophyseal joints.59 Reported complications are uncommon, with rates ranging from 1.5% to 4%. Most complications are device related, including improper sizing, seating, and location of the prosthesis.60 Long-term clinical data are currently lacking, with important questions remaining regarding patient selection, in vivo durability, and revision options.