Dr. Carrino and Dr. Morrison are Assistant Professors of Radiology, Jefferson Medical College, Musculo-skeletal Imaging and Percutaneous Intervention, Department of Radiology, Thomas Jefferson University Hospital, Philadelphia, PA.

Few procedures have generated as much controversy as discography. Discography is controversial not because of the technical aspects or periprocedural complications, but rather because of the questionable downstream decision-making relevance and lack of a criterion standard. There is anatomic evidence and hence concept validity that the disc can be a source of pain (nociceptor) because of the innervation that exists (best established in the lumbar region) along the outer anulus from the ventral nerve roots that provide branches anteriorly (grey ramus communicans) and posteriorly (sinuvertebral nerve). ¹ However, there are many other structures in and around the spine that may be nociceptors, and it is often difficult for the clinician to differentiate these potential sources of pain (or when multiple, which is the primary inciting source), especially if there are numerous imaging "abnormalities." The numerous pain sources have a variety of clinical expressions, which overlap with each other and with other disorders as well.

While the concept of discogenic pain represents a reasonable paradigm, poorly performed discography can assuage the importance of making this diagnosis and has contributed to its dubious reputation. Historically, discography has been used to detect anular tears and disc herniations often as a complement to myelography. After the emergence of cross-sectional imaging techniques (eg, CT and MR) the use dropped off. Currently, the primary purpose for discography is for documentation of the disc as a pain source. ^2,3

The intervertebral disc (IVD) is a composite structure consisting of three distinct components: the nucleus pulposus (NP), the anulus fibrosus (AF), and the cartilaginous endplates. Decreased tissue cellularity and altered matrix architecture characterize intervertebral disc degeneration. The disc derives its structural properties largely through its ability to attract and retain water. The AF is the main torque converter in the spine while the NP provides hydrostatic pressure. Delamination of the anulus is the key patho-etiologic feature that produces a herniated nucleus pulposus (HNP). Rotary strain sets the stage for herniation. Overt trauma has a variable and questionable role but may be the precipitating event superimposed on underlying degeneration. Collectively, these features can lead to abnormal spine biomechanics and pain. Degenerated discs are thought to cause pain in several ways, including mechanical instability (stretching of pain fibers), compressive impingement on adjacent nerves (radiculopathy), and biochemical irritation via the release of inflammation mediators such as phospholipase A2, causing primary dural pain. ⁴

Internal disc disruption (IDD) is a term that was coined in the 1970s to describe pathologic changes of the internal structure of the disc. Internal disc disruption and degeneration involve a physiochemical change in the glycosaminoglycans of the NP, which act to bind water; over time this water-binding capacity diminishes. Disc degeneration is usually heralded by loss of hydration and thus decreased T2 signal on MR imaging. However, focal T2 bright areas reflecting anular tears indicate fragmentation of the outer collagenous AF. Hyper-intense zone (HIZ) is the term that has been coined to denote this finding on T2-weighted MR images. In the patient population undergoing MR imaging for lumbar back pain, this finding may be noted in approximately 25%. The presence of an HIZ correlates with an anular tear and an approximately 85% chance that there will be concordant pain reproduction at discography. ⁵ An HIZ may enhance after contrast administration reflecting the fibrovascular ingrowth into the region of the anular tear (Figure 1). In addition, nerve tissue has also been seen by histology in this lesion and is the purported mechanism by which peripheral anular tears generate pain. The prognostic or therapeutic significance of this finding has not yet been elucidated and asymptomatic HIZs may also be encountered.

The cervical spine is more biomechanically challenged than the rest of the spine. The ligaments and supporting soft tissues are important for mobility. Recent anatomical re-evaluation of this area has determined that the AF is a crescentic anterior interosseous ligament rather than a completely circumferential "o-ring" that surrounds the NP, as in the lumbar spine. ⁶ It tapers laterally where the uncinate processes exist and is deficient in the posterolateral aspects. Posteriorly, there is a thin layer of vertically oriented fibers reinforced by the posterior longitudinal ligament (PLL). This anatomical structure is not present in the human fetus, child, or adolescent, and it is believed to represent a normal phenomenon of maturation. When bipeds turn their heads, there is a rotational component involved; in quadrapeds, this is mostly achieved with lateral side bending. Therefore, uncovertebral joints are unique to vertebrate species that maintain an erect posture, and this biomechanical condition causes uncovertebral hypertrophy as a normal aging (presumably degenerative) phenomenon. In the cervical spine, intradural connections between adjacent nerves may account for the greater than expected overlap of dermatomal pain patterns in this region. Because of these considerations, chronic cervical spine pain of an axial nature is difficult to evaluate and treat.

The thoracic spine is stabilized by the ribs and has less range of motion than the other segments of the spinal column. Thoracic pain is relatively uncommon. However, it is important from a management perspective because dorsal back pain can be as disabling as cervical and lumbar pain. ⁷ While histological studies of the thoracic discs are currently being re-evaluated, ⁸ it has been revealed that branches of the rami communicantes provide innervation circumferentially. ⁷ MR imaging reveals that a substantial number (11% to 12.5%) of asymptomatic degenerative or protruded discs also exist in the thoracic spine. ⁹ However, anatomical changes on imaging studies do not necessarily equate with pain generation. In one investigation, approximately one-quarter of the discs injected provoked a pain response that did not match MR imaging findings or morphologic findings at discography. ¹⁰ One case series on thoracic discography concluded that useful information is obtained for treatment planning. ¹¹ In this study, in addition to painful segments, control discs were also injected, which did not provoke pain.

Low back pain is one of the most common medical problems encountered by healthcare providers. Accordingly, the lumbar spine is the most commonly requested site for discography. For patients whose symptomatology is predominately axial and nonmyelopathic and/or nonradicular, imaging may be insufficient or equivocal for determining the nature, location, and extent of symptomatic pathology. Conversely, imaging reveals asymptomatic abnormalities in a substantial proportion of patients. ¹²

A major controversial issue regarding discography is how effective spine arthrodesis is for alleviating primarily axial pain. Most of the recent literature supports the use of discography in select patients. In general, the role of surgery for axial pain is limited. At present, there is paucity of prospective, randomized or controlled trials evaluating spinal fusion outcomes. However, one study supports that discogenic pain syndromes can be treated by arthrodesis, with a 46% satisfactory outcome. ¹³ Another controversial issue is who should perform discography. Some contend that it should be an individual who is familiar with the patient (ie, the treating physician) and thus may better put the information in context. However, others support that someone other than the operating surgeon should perform discography so that a more objective test is achieved. In either situation, it should be an individual with experience in positioning of needles and with knowledge of fluoroscopic anatomy and spinal pain syndromes (ie, "spine specialist").

Demand for discography is increasing, as a diagnostic tool to determine levels of pain generation for patients who are being considered for surgical management (eg, interbody arthrodesis) or other types of procedures. Degenerated discs may be relatively motionless, and the source of pain may be at the relatively normal-appearing (or at least less-degenerated appearing) levels above or below due to abnormal biomechanics at these levels. Surgeons concerned with limiting the extent of fusion are interested in obtaining more evidence beyond MR imaging abnormalities to document what IVD levels are contributing to the painful syndrome.

Interpretation of a discogram includes a morphologic and functional evaluation. The functional evaluation is more important because MR imaging is well suited for characterization of morphologic findings. The tenet of discography is that injection into the discs and subsequent increased intradiscal pressure will elicit a concordant pain response (one that mimics the patient's typical pain) if that disc is a pain generator. A scale of subjective pain severity from 0 (no pain) to 10 (maximal pain) can be determined during the procedure by asking the patient to relate what their level of pain is during each injection. The patient is also asked whether the pain mimics his/her typical pain (ie, "concordant"). In order to evaluate the patient's pain response more objectively, multiple vertebral levels around the suspected pain generator are injected during the procedure; the patient is not told which level is being injected, or when the injection is starting. In our practice, we coach the patient prior to the procedure regarding reporting of pain response and monitor for spontaneous pain elicited during the examination. It is important to establish a "reference level," or relatively pain-free level with injection. For discography to be considered positive, there should be at least one reference level, which is defined by the absence of pain or lack of concordant symptoms upon injection. An unquestionably positive discogram consists of a single concordantly symptomatic IVD with control discs above and below that level (if it is not the lumbosacral junction). Optimal benefit results when one or two levels demonstrate a highly concordant pain response, with a relatively pain-free adjacent reference level(s). If all levels are painful, a limited fusion may not result in patient satisfaction, and these results can suggest that continued medical management might be the best course, rather than surgery.

A position statement regarding lumbar discography from the North American Spine Society (NASS) was published in 1995. ¹⁴ Specific indications include patients with persistent pain in whom noninvasive imaging and other tests have not provided sufficient diagnostic information. In pre-operative patients who are to undergo fusion, discography can be used to determine if discs within the proposed fusion segment are symptomatic and if the adjacent discs are normal. In postoperative patients who continue to experience significant pain, discography can be used to assist in differentiating between postoperative scar and recurrent disc herniation; or to evaluate segments adjacent to the arthrodesis. Discography can also be used to confirm a contained disc herniation or IDD as a prelude to minimally invasive discectomy or intradiscal therapy.

Discography is performed on an outpatient basis. Guidance for needle placement should be performed with a C-arm, floating image intensifier, or with biplane fluoroscopy. Patients must be informed ahead of time that the purpose of the procedure is to generate a pain response, which in some circumstances can be severe. Complications include persistent pain, infection, bleeding, and injury to exiting nerve roots. ^2,3,15-17 To minimize the risk of disc infection, the procedure should be performed with a surgical-type prep and drape of the patient and surgical scrub, gown, mask, and gloves for the physician. Antibiotics should always be administered, whether intravenous or intradiscal.

Intravenous anesthetic can potentially blunt a positive response and is not necessary for lumbar procedures, since placement of the needles can be performed relatively painlessly with proper technique. A short-acting agent can be used for cervical and thoracic procedures, but must "wear off" or be reversed before injection of the discs. Patients are monitored routinely during the procedure with pulse oximetry and a blood-pressure cuff. Information assessed and recorded should include the volume of contrast injected, pain response with particular emphasis on its location and concordance to clinical symptomatology, and the pattern of contrast distribution. CT imaging can be performed following the procedure if additional information about the location of anular pathology is desired. In our practice, we use CT imaging routinely for cervical and thoracic levels. We do not use it routinely for the lumbar spine but it is becoming more common as a complement to the fluoroscopic images or in order to better delineate and characterize IDD prior to intradiscal therapy.

Preprocedure instructions are similar to other spinal injections or vertebral biopsy-type procedures. These are related to the patient on an instruction sheet as follows. No solid food should be eaten 6 hours prior to the procedure (sips of water for medications are allowed). No aspirin-containing products should be used for at least 1 week prior to the procedure. Nonsteriodal anti-inflammatory medications (eg, acetaminophen or ibuprofen) or other pain control medication is acceptable as long as it does not contain any aspirin. However, all pain medication should be discontinued on the day of the procedure. If the patient is a diabetic taking insulin, the patient should consult his or her primary physician regarding the insulin dose to take the morning of the procedure. Patients may continue blood pressure medication unless contraindicated by their physicians. In general, it is recommended that patients review all medications with their primary physicians no later than 3 days prior to their procedure. The patient should bring any relevant outside imaging studies. The patient will need to rest in the recovery area for 60 to 90 minutes following the spinal injection. Patients must have a companion for discharge after the procedure. We also provide a telephone number to contact the radiologist or support staff if need be.

Postprocedure instructions for the patient include surveillance for signs and symptoms of disc infection. Discitis following either lumbar or cervical discography is an uncommon occurrence. ^18,19 It is, however, debilitating for the patient and can pose a diagnostic dilemma for the physician. Signs and symptoms are not always clear and the diagnosis is often delayed secondary to inconclusive laboratory and imaging studies early in the course of the illness. An elevated erythrocyte sedimentation rate and narrowing of the disc space have been noted after several days of pain, but these are often normal at the initial presentation. Discitis occurs in 1% to 4% of patients undergoing discography; however, the frequency can be minimized by prophylactic antibiotic administration. For cervical and thoracic discography, intravenous antibiotic prophylaxis is preferred (cephazolin or an equivalent cephalosporin). For lumbar discography, a cephalosporin antibiotic can be mixed with the contrast material injected. In our practice, we reconstitute 1 gram of cephazolin with 10 mL of nonionic contrast suitable for intrathecal administration. Preliminary data show that uncomplicated discography does not produce MR imaging abnormalities immediately or within the 2 to 6 week period following intradiscal injection. ^20,21 Therefore, MR imaging is suitable for evaluation of potential postdiscography complications.

Specific technical issues

Cervical spine technique

Cervical discography is performed using an anterior approach. Since the complication and false-positive rates appear to be higher than those for lumbar discography, cervical discography is performed less frequently than lumbar discography.

The patient is placed in the supine position with a small rolled towel between the scapulae to extend the neck and a small pillow under the neck itself for comfort. Disc puncture is usually accomplished using anteroposterior imaging for frontal visualization (Figure 2). The skin of the anterior and anterolateral neck from the level of the mandible to the supraclavicular region is prepped. The esophagus lies to the left of the spine at the level of C7 in most individuals. Therefore a right-sided approach is normally used, especially for right-handed operators. Firm but gentle pressure is applied to the space between the trachea and medial border of the sternocleidomastoid muscle, displacing the laryngotracheal structures to the left. The right carotid artery is maintained underneath the fingers. With this maneuver, the anterior surface of the spine can be palpated in almost every individual. The needle entry point should be adjacent to the medial border of the sternocleidomastoid muscle but not through the muscle belly. Using this landmark, the skin puncture site will be more lateral for the cephalad disc levels and more medial as one progresses caudally. Initially, the needle is directed to the vertebral body just below the endplate in order to ascertain the depth. Subsequently, minimal retraction and cephalad migration will direct the needle onto the anterolateral surface of the disc anulus and with small incremental movements can be advanced into the center of the disc (Figure 3). The adult cervical disc normally accepts a volume of <0.5 mL. Usually a 25-gauge single needle approach will suffice. Skin and periosteal anesthesia is often unnecessary and may confuse the interpretation.

As mentioned, the lateral and posterolateral portions of the cervical disc anulus are relatively attenuated. This results in clefts (joints of Lushka) that communicate with the nucleus, which are unique to the cervical spine. Opacification of these regions in patients older than 20 years of age should not be confounded for degenerative disc disease based on this morphologic finding alone (Figure 4).

Thoracic spine technique

Thoracic spine discography can be performed in the prone semi-oblique 45š position (using a wedge) with the less painful side up. Alternatively, the patient may be placed prone and anteroposterior images obtained with the end plates in alignment. The C-arm is rotated to the side of injection until a lucent zone directly in line with the beam is seen projecting over the thoracic disc (Figure 5). This usually requires approximately 20š of rotation. The needle should enter the disc lateral to the interpedicular line and medial to the costovertebral joints in order to avoid potential complications, such as accidental puncture of the lung or thecal sac. Similar to the C-spine, this is done as a single-needle technique. Usually 25-gauge needles will suffice for small individuals; however, 3.5-inch, 23-gauge needles are often preferred because they are stiffer and can negotiate better around nerve roots and/or the osseous structures if necessary. Also similar to the cervical spine, the thoracic disc normally accepts a small volume of injectant (<1.0 mL). Fluoroscopic images may be difficult to interpret because of the superimposition of osseous structures, difficulty obtaining a true lateral projection, and presence of a small amount of injectant (Figure 6). Therefore, post-discography CT imaging is often a useful adjunct to delineate IDDs and HNPs (Figure 7).

Lumbar spine technique

While the original description of lumbar discography used a transdural midline approach, currently most operators use a posterolateral extradural approach. The technique for lumbar injection is as follows (Figure 8). Levels for injection are chosen based on imaging findings, clinical examination, and surgical options, and should be discussed with the referring physician prior to performing the procedure. Injection generally includes L3-4, L4-5, and L5-S1. The patient is positioned in a prone-oblique position with the less painful side up. Each level is set up fluoroscopically so the disc is parallel to the beam and obliqued so that the superior articular process of the overlying facet joint is slightly posterior to the center of the endplate (30% to 50% zone). Lidocaine is administered under the skin. Next a 22- or 23-gauge 3.5-inch needle is advanced along the X-ray beam toward the disc, past the anterior margin of the superior articular process. Anesthetic is injected as the needle is withdrawn, thus fully anesthetizing the path to the disc. Care must be exercised with respect to the depth of the injection, so as not to inject the anulus or nerve root sheath.

A coaxial technique is used to place the discography needles into each IVD. This reduces trauma to the anulus, and may reduce the risk of infection. The larger outer needle allows rapid positioning at the disc margin, with a small- gauge needle used to penetrate the anular fibers. At the L3-4 and L4-5 levels, a 20-gauge 3.5-inch outer needle can be used in conjunction with a 6- to 8-inch 25-gauge inner needle. The L5-S1 disc may be located below the pelvic rim and can be difficult to access. Generally, the X-ray beam is oriented with more caudal angulation than the higher levels and is rotated slightly to open a small triangle of access over the iliac crest (Figure 9). When this window is achieved, one needs to determine if the lumbosacral IVD can be punctured in the central portion with a direct approach or if the orientation is more parasagittal. If the course of the outer needle is parasagittal, a curved-needle technique is required to position the inner needle centrally. Prior to inserting the inner needle, a bend (or curve) may be applied along the distal tip so that when it emerges from the guide needle, the inner needle deflects toward the center of the disc. Alternatively, pre-curved coaxial needle systems are available. Typically, a 22-gauge inner needle through an 18-gauge outer needle is used for most cases that require a curve. In addition, particularly at the lumbosacral junction, longer needles may be required.

Positioning of all needles during placement is checked frequently in the plane along the trajectory of the needle and is supplemented with the anteroposterior and lateral planes as the tip approximates the disc. The tip of the inner needle should be positioned as close as possible to the center of the disc, so that injection is into the nucleus pulposus (Figure 10) instead of the innervated annular fibers, which can result in a false-positive pain response. After all needles are placed, 1 to 2 mL of contrast (mixed with antibiotic) is injected at each level, with fluoroscopic monitoring and evaluation of any pain elicited. A morphologically normal disc demonstrates a central globule of contrast collection or "hamburger-bun" configuration and degeneration is indicated by a horizontal, linear distribution of contrast (Figure 11). An anular tear is diagnosed if contrast extends into the periphery of the disc in the expected region of the AF (Figure 12). CT imaging may be used to complement projectional imaging techniques, and grading systems are available to characterize IDD. ³

Future considerations and conclusion

There is an interest in characterizing and segmenting patients based on the results of pressure-controlled manometric discography. This technique may help stratify patients into categories that are more likely to improve from interbody fusions. ²² However, there are issues regarding whether intradiscal injection, which produces a tensile load is comparable pathophysiologically to the compressive load that is exerted by virtue of our bipedal existence. Diagnostic blocks are sometimes used in other areas to identify nociceptive sites. Therefore, rather than provoke pain by recapitulating pressure, another route being explored is an attempt to diagnose discogenic pain by sinuvertebral nerve blocks as an alternative to provocative discography. ²³ While the diagnostic utility of discography is quite evident, the treatment utility based on the patient outcome is paramount. Therefore, the value-added feature that discography should provide is to identify patients amenable to available therapies and not to contribute to the management dilemma. Meanwhile, less invasive forms of intradiscal therapy are also evolving (eg, percutaneous disc decompression using coblation and intradiscal electrothermal anuloplasty), which may make discography more relevant. One important consideration is that biomechanical and biological factors constrain the path of disc repair. Thus, biomaterials that either provide scaffolding or promote healing are theoretically favorable, and in order to be deployed successfully, they will need a suitable percutaneous delivery system. Therefore, discography techniques will likely be important for the next generation of minimally invasive intradiscal therapies. AR

Disclaimer: This article is an instructional and educational vehicle that complements, but does not replace, the time, proctoring, and practical experience needed to perform discography. Familiarity with vertebral body bone biopsy techniques and basic spinal injection principles are valuable prerequisites. A hands-on workshop with cadavers or mini-fellowship is strongly recommended.