INTRODUCTION

Discography was developed in the late 1940s for diagnosing lumbar intervertebral disc herniation.^1,2 In contemporary practice, discography refers to provocation discography in which the most important component is the evaluation of pain reproduction caused by pressurizing the disc with contrast medium. Discography is conceptually an extension of clinical examination, tantamount to palpating for tenderness.³ A precision injection of contrast dye into the disc nucleus stimulates nerve endings⁴ (via two mechanisms: chemical stimulus from contact between contrast dye and sensitized tissues, and a mechanical stimulus resulting from fluid-distending stress).

Discography is a potential solution to the diagnostic dilemma concerning which patients to treat surgically and at what segmental level. In 1995, the North American Spine Society stated that discography, especially followed by CT scanning, may be the only study capable of providing a diagnosis or permitting precise description of the internal anatomy of a disc and the integrity of disc substructures.⁵ Although new diagnostic imaging tools have been developed and are widely used, discography is still practiced. Discography remains particularly useful in problematic cases unresolved by MRI or myelography and in patients for whom surgery is contemplated.⁶

In this chapter, the technical considerations and complications of discography are discussed.

INDICATIONS AND CONTRAINDICATIONS

According to the position statement on discography by the North American Spine Society:⁵

The single purpose of discography is to obtain information. Disc morphology is not diagnostic; it does not indicate whether a given disc is responsible for a patient’s pain. However, it is reassuring if a disc proven to be symptomatic on other grounds also happens to be morphologically abnormal. The prime indication for discography is to establish a diagnosis of discogenic pain when therapy is to be directed at that disc. A parallel application is to identify asymptomatic discs. When a single disc is found to be symptomatic in the presence of adjacent asymptomatic discs, focused surgical therapy can be entertained. Patients with symptomatic or abnormal discs at multiple levels constitute a greater surgical challenge.

Congenital anomalies of the vertebrae or nerve roots and postoperative spinal abnormalities constitute relative contraindications to discography. Such conditions require greater care and dexterity to negotiate the abnormal anatomy to access the target disc and avoid injury to surrounding structures.

In the case of cervical discography, spinal cord compression constitutes an absolute contraindication to discography. In such cases neurologic features are key, and information about discogenic pain is immaterial to management. Moreover, the uncertainty of the relationship between the prolapsed disc and the spinal cord and resistance of the disc to passage of a needle only invites morbidity.

PREPROCEDURAL EVALUATION AND PATIENT PREPARATION

TECHNIQUE OF LUMBAR DISCOGRAPHY

Disc puncture

Prior to injection, a fluoroscopic examination of the spine is performed to confirm segmentation and determine the appropriate level for needle placement. Using AP view, the beam should be parallel to the inferior vertebral endplate. After selecting the target disc using AP view, the fluoroscopic beam is axially rotated until the facet joint space is located midway between the anterior and posterior vertebral margins. In this view, the insertion point is 1 mm lateral to the lateral margin of the superior articular process (Fig. 25.1).

Fig. 25.1 Oblique view of lumbar spine (*, needle insertion point; P, pedicle; FJ, facet joint).

The insertion point is marked on the skin. Since the distance between the opposite superior articular processes increases at lower levels, the usual distance from the midline increases from about 3–4 cm at the T12–L1 level to 6–7 cm at the L5–S1 level. Because of the iliac crest and increased interfacetal distance, at the L5–S1 level the fluoroscopy tube is rotated only far enough to bring the facet joint space approximately 25% of the distance between the anterior and posterior vertebral margins.

Prior to needle placement, the skin, subcutaneous tissues, and deep muscular tissues along the needle trajectory are infiltrated with local anesthetic (1% lidocaine). Techniques vary depending upon the number of needles used. The single-needle technique has the advantage of a single needle insertion and perhaps shorter procedure duration, but the risk of disc infection may be higher.

To avoid potential neural injury, the needle should be directed into the safe triangle. The borders of the safe triangle include the nerve root for the superior tangential border, the vertebral endplate of the target disc for inferior border, and the lateral margin of the superior articular process for the medial side line (Fig. 25.2). To minimize nerve trauma, one should use a needle with a short, noncutting bevel or blunt-pointed tip with a side port.

Fig. 25.2 Safe triangle for needle insertion (NR, nerve root outlined by contrast media; SAP, superior articular process; EP, endplate).

Puncture of T12–L1 through L4–L5 intervertebral discs

In the double-needle technique, a styletted 25-gauge, 6-inch needle is placed into each disc through a 20-gauge 3.5-inch introducer needle under fluoroscopic guidance. To protect the discographer’s hand from radiation exposure, forceps may be used to grasp the introducing needle. The introducer needle is advanced parallel to the fluoroscopic beam using an oblique fluoroscope view (Fig. 25.3).

Fig. 25.3 The introducing needle should proceed parallel to the fluoroscopic beam using an oblique fluoroscope view.

A slight ‘hockey-stick’ bend at the end of the introducer needle can improve navigation. If bony obstruction is encountered, the physician should confirm whether the needle has contacted the superior articular process or the vertebral body. If necessary, the needle may be slightly withdrawn and its trajectory modified. The introducer needle can be either advanced just over the lateral edge of the superior articular process or advanced to the margin of the disc. At the L5–S1 level, advancement may proceed just over the lateral edge of the superior articular process. When the introducer or discogram needle contacts the disc margin, the ideal position in the AP projection is on a line drawn between the midpoints of the pedicles above and below (Fig. 25.4A).

Fig. 25.4 The ideal position of needle when the needle contacts the disc is on the line between midpoint of pedicles at AP (A) and posterior vertebral margin at lateral view (B) (P, pedicle).

In no case should one advance the introducer or discogram needle medial to the inner pedicle margins before contacting the intervertebral disc. In the lateral view the needle should contact the disc between the posterior vertebral margins (Fig. 25.4B).

After confirming introducer needle position with a lateral view, a 25-gauge, 6-inch needle is advanced into the center of the disc through the introducing needle while monitoring by lateral view. A slight ‘hockey-stick’ bend on the end of the discogram needle facilitates navigation. When the needle contacts the disc, position should be checked using AP and lateral views. Contact with the anulus fibrosus is characterized by the perception of firm but resilient resistance, and frequently the patient experiences a momentary, sharp, or sudden aching sensation in the back or the buttock. The needle is then advanced to the center of the disc. Needle position must be monitored and checked by both AP and lateral imaging.

L5–S1 intervertebral disc

Due to increased facetal width and the presence of the iliac crest diacrest, puncture of the lumbosacral disc is more challenging. Instead of a direct lateral approach, once the introducer needle is advanced to the anterior border of the superior articular process of S1, a slight curve or ‘hockey-stick’ bend is needed to advance the discogram needle in a medial and slightly posterior direction around the SAP to contact the disc just anterior to the vertebral margin as viewed in the lateral fluoroscopy projection. Less experienced operators may find an 18/23-gauge needle combination easier to direct than the 20/25-gauge combination used for upper levels. Longer needle combinations may be required in muscular or obese patients. The fluoroscopy tube is rotated until about 1–2 cm of the L5–S1 disc is visualized between the superior articular process of S1 and the sacral ala (Fig. 25.2).

Using the oblique fluoroscopic view, the guide needle is introduced toward the bony notch between the sacral ala and superior articular process of S1 until the needle tip lies immediately adjacent to the anterolateral aspect of the superior articular process of the sacrum (Fig. 25.5A). The needle tip should not be handled directly, but should be wrapped in sterile gauze. The distal 2–3 cm of the needle should be bent in a direction opposite the bevel. The degree of curve is determined by the operator on the basis of how much deflection is required in the patient at hand for the needle to approach the target disc center. In a lateral fluoroscopic projection, the 25-gauge discogram needle is passed through the guide needle while the guide needle is held firmly in position. The inner needle is advanced until the tip emerges from the guide needle. The inner needle is advanced under direct fluoroscopic vision. As it emerges, the guide needle is retracted slightly (Fig. 25.5A). This unsheathes the procedure needle, which should be turned so that the curve or bend bows the introducer needle in a medial and posterior direction through the safe triangle. Once the needle encounters the anulus fibrosus, its position is checked and confirmed in both AP and lateral fluoroscopy views. In the lateral view the needle should contact the disc 2–3 mm anterior to the vertebral margin (Fig. 25.5B) and in the AP view the needle should ideally be on a line bisecting the midpoint of the L5 and S1 pedicles.

Fig. 25.5 (A) Oblique view of L5–S1 discography. The inner needle is advanced slightly under direct fluoroscopic vision, and the guide needle is simultaneously retracted slightly. (B) Lateral view. The inner needle should contact the disc 2–3 mm anterior to the vertebral margin.

If the inner needle fails to curve medially it will not pass toward the center of the disc, and may strike the ventral ramus. As a result, the needle course must be monitored. If the needle fails to track medially and posteriorly, the needle should be removed and its curvature accentuated. Should the procedure needle meet with bony obstruction, the fluoroscope should be turned to determine whether the superior articular process or the vertebral body has been encountered. If the vertebral body has been encountered, the course of the needle can be corrected by withdrawing it slightly and rotating the needle appropriately. If the needle is blocked by the superior articular process, the inner needle is retracted into the guide needle and the pair are advanced to the lateral edge of the S1 superior articular process. The inner needle may then be directed toward the disc as described above. The ideal final needle position is in the disc center; however, there is leeway. In severely degenerated discs the needle position is not as critical, since contrast medium will spread throughout the disc. Ideally, the needle should be within 4–5 mm of the center on AP and lateral fluoroscopy.

Provocation using pressure manometry

Provocation

Once the needle tip has been properly placed in the center of the nucleus pulposus, nonionic contrast medium mixed with antibiotic is injected into each disc at slow velocity using a controlled injection syringe with digital pressure readout. The total volume injected should probably be limited to 3.5 mL. Although a few severely degenerated discs will accept more volume, the incidence of false-positive pain responses may increase. If one cannot achieve 50 psi above opening pressure at 3.5 mL due to the degree of disc degeneration or a leak through the endplates or anulus, a dynamic pressure of 50 psi above opening can usually be achieved by more rapid injection. In general, however, the media should be injected as slowly as possible. Higher injection speeds may cause rapid pressure elevations leading to increased pressure differences between the nucleus pulposus and manometer and between the dynamic and static pressures. At 0.05 mL/sec (one revolution of the Merit monometer) the difference between static and dynamic pressures is minimal. The intrinsic disc pressure is the pressure required to start the flow of contrast medium into the nucleus (opening pressure) resulting from osmotic forces within the disc resisted by anulus tension and the tension of the anterior and posterior longitudinal ligaments. Typical opening pressures are 5–25 psi, depending on the degree of nuclear degeneration. An opening pressure >30 psi usually indicates that the needle tip is within the inner anulus. In this case, the needle tip should be repositioned.

The disc is slowly pressurized by injecting 0.5 mL increments through a syringe attached to a pressure measuring device. At each 0.5 mL increment, the injection pressure, location of contrast medium, and any pain response are recorded. At a slow injection speed, the dynamic pressure more closely reflects the real intradiscal pressure than the static pressure taken without pressure applied to the injecting syringe. Injection continues until one of the following end points is reached: subject pain = 6/10, intradiscal pressure >50 psi above opening in a disc with a grade 3 annular tear or 80–100 psi in a normal-appearing nucleogram, or a total of 3.5 mL of contrast medium has been injected.

Imaging

The appearance of the normal nucleus following the injection of contrast medium is unmistakable: the contrast medium assumes a lobular pattern or a bilobed ‘hamburger’ pattern (Fig. 25.6A). A variety of patterns may occur in abnormal discs.⁹ Contrast medium may extend into radial fissures of various lengths but remain contained within the disc (Fig. 25.6B), or it may escape into the epidural spaces through a torn anulus (Fig. 25.6A). In some cases (Fig. 25.6C), the contrast medium may escape through a defect in the vertebral endplate.³ However, none of these patterns alone is indicative of whether the disc is painful; that can be ascertained only by the patient’s subjective response to disc injection.

Fig. 25.6 The appearance of the nucleus following the injection of contrast media. (A) The contrast medium assumes either a lobular pattern or a bilobed (‘hamburger’) pattern (black arrowhead) as a normal nucleogram. Contrast medium escaped into the epidural space through a radial fissure at L5–S1 level (white arrow). (B) Contrast medium extended into radial fissures (arrow) but remain contained within the disc. (C) The contrast medium escaped through a defect in the vertebral end plate (white arrowheads).

Immediately after discography, CT–discography may be performed to define fissures extending to the outer third of the anulus and extending circumferentially within the anulus fibrosus. Radial annular tears can be found by discography, but only the postdiscogram CT axial view clearly shows the location and size of fissures within the anulus fibrosus.³ Sachs et al.¹⁰ developed the Dallas discogram scale, in which annular disruption is graded on a 4-point scale. Grade 0 describes contrast medium contained wholly within a regular nucleus pulposus. Grades 1 to 3 describe the extension of contrast medium along radial fissures into the inner third, middle third, and outer third of the anulus fibrosus, respectively. Aprill and Bogduk¹¹ proposed a modified Dallas description scale which includes grade 4, distinguished from grade 3 by the spread of contrast medium circumferentially within the substance of the anulus fibrosus and subtending a >30° arc at the disc center.

Interpretation

The most important information obtained from discography is whether the patient’s pain is reproduced. There is no alternative or superior means of determining if a disc is the source of a patient’s pain. Conceptually, discography is an extension of clinical examination, tantamount to palpating for tenderness. It is only the inaccessibility of a disc to palpation that requires the use of needles. In this regard, it is critical that the criteria for a painful disc be rigorously satisfied. Internal control observations are mandatory: a disc cannot be deemed the source of a patient’s pain if stimulating other discs or other structures in the same region reproduce similar pain.

Assessing patient response to discography requires measuring the pain reproduced by injection. Pain may be characterized by three components: intensity, location, and character. If the location and character of pain provoked during discography are similar to or the same as the patient’s clinical symptoms, the criteria for concordant pain are satisfied. The intensity of pain is measured by the patient (e.g., by using a numerical rating scale) and by observed pain behaviors. The intensity of provoked pain, however, is dependent on stimulus intensity. In simple terms, the harder one pushes on the syringe the more likely the disc is to hurt. By measuring intradiscal pressures, stimulus intensity can be quantified and standardized, permitting more reliable comparisons among patients and discographers. While injection pressures may be manually estimated, use of a controlled inflation syringe with digital pressure readout is more precise.

Adding pressure monitoring to provocative discography improves interobserver reliability and, as a result, reproducibility. A positive discogram requires an abnormal disc, pain response =6/10 (numeric rating scale; NRS), pressure level =50 psi, pain described by the participant as ‘familiar,’ and at least one negative control disc. One must, however, be aware that transient pain is often provoked when fissures are suddenly opened. In most cases, such pain should not be used as evidence of a true-positive response unless pain >6/10 is sustained for more than 30 seconds. Most experienced discographers will perform a confirmatory re-pressurization once a suspected positive response is provoked. If re-pressurization does not again provoke significant concordant pain at 50 psi or less above opening pressure, then the initial response will remain indeterminate. These criteria for positive response were reconfirmed in a study performed in 13 normal asymptomatic volunteers.^11a,11b When the operational criteria for discography were set to pressure =50 psi and evoked pain intensity >4, the expected false-positive rate was <10%. However, a false-positive rate of zero could be secured either if the pain score was held at 4 and the injection pressure lowered to 20 psi, or if the pressure was held at 50 psi and the required pain score was raised to 6. To increase discography specificity, local anesthetic may be injected into the positive disc in an effort to obtain prolonged relief of pain from that disc (analgesic discography). In a preliminary study investigating the reliability of analgesic discography, 78% of patients showed significantly prolonged pain relief after local anesthetic injection. Those patients underwent fusion surgery and were followed to observe surgical outcome (T. Alamin, personal communication, 2004).

In addition to these criteria for positive discography, disc anulus sensitivity may also be graded. Using the protocol of Derby et al.,¹² four disc categories may be defined: (1) chemical discs, which provoke pain at 15 psi above opening pressure, (2) mechanical discs, which provoke pain at pressures between standing and lying, or 15–50 psi above opening pressure, (3) indeterminate discs, which provoke pain at 51–90 psi above opening pressure, and (4) normal discs, with no pain provocation. If a disc is painful at >50 psi, the response cannot be considered clinically significant, since it is difficult to distinguish from the effect of mechanically stimulating a normal or subclinically symptomatic disc.¹³ Excessive stimulation involving pressures >50 psi above opening pressure and uncontrolled, high injection speeds increase false-positive responses.

TECHNIQUE OF CERVICAL DISCOGRAPHY

Disc puncture

Midline approach

The disc level to be studied is identified by fluoroscopy. Most discographers use the AP view and count both upwards from the C7–T1 level and downwards from the C2–3 level. Counting down from the C2–3 level using the lateral fluoroscopic view is more accurate. The lateral views of lower cervical segments are usually attenuated by the overlying shoulders. To optimally visualize these segments, longitudinal, downward traction of the bilateral upper limb is often necessary.

Once the levels are identified, the fluoroscopy tube is rotated in a cephalad–caudal direction to bring the endplates parallel to the beam. Pressure is applied with the index finger to the space between the trachea and the medial border of the sternocleidomastoid. Firm but gentle pressure will displace the great vessels laterally and the laryngeal structures and trachea medially. Below C4, the right common carotid artery and the internal carotid artery above C4 artery are palpated. The fingers are insinuated until they encounter the anterior surface of the vertebral column. The patient is monitored closely for vasovagal signs, which may be caused by compression of the catrotid artery during manual displacement with needle entry.¹⁴

The spinal needle or metal instrument can be used as a guide for needle insertion. The instrument is placed on the skin parallel to the disc space using fluoroscopic guidance and a marker pen is used to draw the lines on the skin over the intervertebral disc spaces. The needle entry point should be medial to the medial border of the sternocleidomastoid, and not through that muscle. The declination of the sternocleidomastoid ensures that at C3–4 the puncture point lies more laterally and will avoid the pharynx, whereas at C7–T1 it will be more medial and avoid the apex of the lung. One should use the shortest needle possible to reach the center of the target disc. Longer needles are more difficult to hold and direct and easier to inadvertently pass through the disc. A 23- or 25-gauge, 2.5-inch spinal needle is typically used for the procedure. With the point of the needle just medial to or under the index finger, both the needle and index finger can be moved in unison. The trachea is pushed medially by the fingernail of the index finger, and when the needle overlies the disc at a 20–40° angle, the needle is introduced through the skin directed toward the anterior lateral aspect of the disc (Fig. 25.7). In some patients, the discs can be directly palpated and the finger moved almost perpendicular to the disc center. Patients with short, thick necks or a large, immobile larynx are a challenge. On contact with the anulus, the patient may experience transient pain. The needle is advanced into the substance of the disc under direct AP fluoroscopic visualization. All movements of the needle should be slow and deliberate. One must be very careful not to pass the needle through the disc and into the epidural space or spinal cord. Use of a short needle may be helpful. Novice discographers may want to touch the anterior disc body just above or below the disc margin to help judge the depth of penetration. Once the needle is passed several millimeters into the disc, the lateral view is used to guide further advancement.

Fig. 25.7 Needle trajectories for midline and lateral approaches. In the midline approach, the needle will pass through the platysma muscle, between carotid sheath and airway, longus colli muscle and intervertebral disc. In the lateral approach, the needle will pass through the lateral neck muscles, posterior to internal jugular vein, 1–2 mm anterior to uncinate process.

Lateral approach

Although in the medial approach the needle traverses the same tissue plane used by surgeons to gain access to the cervical intervertebral disc during open and percutaneous surgical procedures, many nonsurgeon interventionists have been trained to use a more lateral approach. In this approach, the disc level is identified by fluoroscopy, and the AP view is adjusted until the vertebral endplates of the target level show parallel lines on the fluoroscopic image. The fluoroscopic beam is then axially rotated until the anterior margin of the uncinate process is moved approximately one-quarter of the distance between the anterior and posterior lateral vertebral margins. In this view, the target insertion point is 1–2 mm medial to the anterior margin of the uncinate process (Fig. 25.8). The skin entry point will be over the lateral neck muscles and posterior to the great vessels or trachea (Fig. 25.9). Pressure displacement of the great vessels is difficult and usually not done. However, this anatomic region is highly vascular and one should carefully observe patients for signs of hematoma during and after the procedure. A 25-gauge, 2.5-inch spinal needle is directed toward the anterior aspect of the uncinate process and should enter the disc 1–2 mm anterior to the process, then slowly advanced to the center of the disc. One should use a needle no longer than is required to reach the center of the target disc. Before injection of contrast, the needle position within the disc is confirmed with both AP and lateral fluoroscopic images (Fig. 25.10). At C7–T1 the medial approach is preferred to avoid puncturing the apex of the lung.

Fig. 25.8 Needle insertion point. The target insertion point is 1–2 mm medial to the anterior margin of the uncinate process using an oblique view (UP, uncinate process).

Fig. 25.9 For the midline approach, the fingernail of the left index finger applies firm pressure to push the great vessels laterally and the trachea medially. The medial border of the sternocleidomastoid is a relative skin surface marker.

Fig. 25.10 The needle placement. AP (A) and lateral (B) views demonstrating the needle tip (arrowhead) in the center of discs.

Provocation and interpretation

The utility of discography for solving puzzling presentations of atypical pain resulting from cervical lesions has been demonstrated. In such cases, MRI did not disclose the pain-producing lesion.^14,15 Normal cervical discs accept fairly small volumes (0.25–0.5 mL) and intradiscal injection of normal discs is not painful.^15,16 Schellhas et al.¹⁵ studied the accuracy of discography and MRI in the identification of source(s) of cervical discogenic pain. The results showed that discographically normal discs were never painful, significant cervical disc anulus tears often escape MRI detection, and that MRI cannot reliably identify the sources of cervical discogenic pain.¹⁵

Pain questioning and fluoroscopic imaging are performed as in lumbar discography. However, pressure-controlled injections are usually not used. If the tip of the needle has been correctly placed in the center of the disc, a 1 mL or 3 mL syringe containing contrast media is attached. Manual syringe pressure is increased slowly until the intrinsic disc pressure is exceeded. Volumes as small as 0.2 mL may cause visible separation of the vertebrae, which should be monitored by fluoroscopy. Concordancy and pain intensity are recorded at 0.2 mL increments.

Use of an NRS can help the patient quantify the intensity of pain experienced relative to the intensity of pain during the injection of adjacent discs. In healthy adults, the intervertebral discs of the cervical spine have a structure similar to that of the lumbar discs, consisting of an anulus fibrosus and nucleus pulposus. However, it has been observed that in the first and second decades of life before complete ossification occurs, lateral tears occur in the anulus fibrosus. Tears are most probably induced by motion of the cervical spine in the bipedal posture.¹⁷ In childhood, the uncovertebral joints begin to undergo an ongoing transformation.¹⁸ Tears in the lateral part of the disc tend to enlarge toward the medial aspect of the intervertebral disc. The development of such tears, uncinate fissures, through both sides may result in a complete transverse splitting of the disc. Such a process may be observed in the second and third decades of life in the lower cervical spine when the intervertebral disc is split in the middle into equal halves.¹⁷ These uncinate fissures are a normal finding, and rarely will one see a ‘normal’ nucleogram.

A normal cervical disc offers firm resistance, and accepts less than 0.5 mL of solution with little discomfort at the time of distention. In most patients there will be transient pain when the uncinate fissures fill with contrast (Fig. 25.11). Pain provocation should be ignored unless persistent, concordant, and significant pain can be reproduced by repeated gentle manual pressurizations. Approximately 1 mL of injected contrast medium is usually enough to opacify the nucleus and fill lateral or posterior annular fissures. A positive response requires provocation of significant (>6/10) concordant pain during a confirmatory repeat injection of another 0.1–0.2 mL of contrast medium. One should also consider retesting adjacent levels to compare pain intensities and concordance of pain. It is not uncommon to provoke concordant pain at multiple levels. In addition, one often will cause separation of the endplates during pressurization, and this movement may cause pain secondary to a symptomatic z-joint.

Fig. 25.11 Cervical nucleogram. (A) Postinjection AP view. All needles were placed in the center of the discs. Contrast medium extended into uncinate recesses. Extension of contrast media to uncinate recesses are common after adolescence with maturation of cervical disc. (B) Postinjection lateral view.

A convincing, positive response to disc stimulation is one in which the patient reports exact or similar reproduction of pain on stimulation of a given disc, stressing one or two adjacent discs is painless or evokes pain totally foreign to the patient’s previous experience. Any other pattern of response is not held to be reliably indicative that the stimulated disc is the source of the patient’s pain. It is common for a patient to have two symptomatic discs, but under those circumstances, one should still identify an adjacent disc that is asymptomatic. Without an asymptomatic ‘control’ disc, there is no evidence that the patient can discriminate between a symptomatic and asymptomatic discs, and there is no evidence that what is reported as disc pain is not simply the pain of needles felt in the back or the neck. Patients may also simply comply with the operator’s expectation that there should be pain. Although the authors agree in principle, in the common occasion when more than four discs provoke concordant pain, performing additional levels (usually the C2–3 or C7–T1) may be inappropriate.

In addition to provocative testing, analgesic discography may be a useful technique for the evaluation of chronic cervical pain.¹⁴ Patients whose clinical pain pattern is reproduced or provoked during the injection of contrast media may benefit from analgesic discography. Once a painful disc has been identified in the course of discography, and once all discographic radiographs have been taken, 0.5 mL of bupivacaine can be injected into the disc to test pain relief. The patient’s accustomed pain should be relieved by this action. The duration of relief should be monitored and recorded by the patient. Relief of accustomed pain for a period consistent with the expected duration of action of the local anesthetic agent used constitutes strong evidence that the disc in question is the pain source. In practice, however, pain relief following multiple-level provocation discography is difficult to assess. Analgesic discography is probably best evaluated at a separate session, and only the one disc in question should be injected with local anesthetic.

Finally, there are some patients who cannot tolerate the procedure. In these cases, little or no useful information may be obtained except that either the patient has a low pain tolerance or the operator performing the procedure is inept.

TECHNIQUE OF THORACIC DISCOGRAPHY

Disc puncture

The current standard technique of thoracic discography was described by Shellhas et al.¹⁹ in 1994. Prior to the procedure, a fluoroscopic examination of the thoracic spine is performed to confirm segmentation and determine the safest and most accessible route to each disc. After the discographer selects the target disc on AP view, the fluoroscope is rotated in a cephalad–caudal direction to align the endplates. The fluoroscopic beam is then rotated ipsilaterally until the corner of the intervertebral disc space is visualized between the superior articular process and the costovertebral joint. Typically, this degree of ipsilateral rotation will superimpose the tip of the spinous process on the edge of the contralateral vertebral body. In this view, the insertion point is just lateral to the interpedicular line (Fig. 25.12) and approximately 3 cm lateral to the spinous process. Prior to needle placement, the skin and subcutaneous and deep muscular tissues along the trajectory of each needle are infiltrated with local anesthetic. Most discographers prefer a single-needle technique using either a 23- or 25-gauge, 3.5-inch needle. A slight bend placed on the end of the needle will facilitate changing directions by needle rotation.

Fig. 25.12 Target point for needle insertion at thoracic spine (oblique view) (SAP, superior articular process; P, pedicle; CVJ, costovertebral joint; EP, endplate).

The trajectory of the needle is roughly parallel and behind the rib as it passes anterior to attach to the spine at the costovertebral joints. One aims at a round to square section of the posterior lateral disc that can be seen through a 1–3 mm opening between the superior articular process and the rib (Fig. 25.13). The needle should be advanced in short increments and the direction changed as necessary by needle rotation. If one stays behind the rib there is no chance of penetrating the lung. Although passage of the needle behind the rib is usually uneventful, passage of the needle between the rib and superior articular process may be difficult. The needle often contacts either the superior articular process or the body of the rib, and it is often difficult to tell which structure obstructs needle advancement. The bend on the needle can be rotated one way or the other and readvanced to assist a needle direction change and allow passage through the small aperture. Once the needle has passed anterior to the superior articular process using a lateral fluoroscopic view, the needle bend is turned posteriorly to facilitate advancing the needle in a more posterior direction (Fig. 25.13).

Fig. 25.13 The needle placement. AP (A) and oblique (B) views demonstrating the needle tip (arrowhead) in the center of discs.

POSTPROCEDURAL CARE

Patients are carefully examined for vital signs and any evidence of subcutaneous bleeding or hematoma immediately after discogram, and again 30 minutes later. After the second postprocedure check, patients are advised of any possible side effects to watch for during the next several days and are allowed to leave the department on the same day. The patient should be warned to expect postprocedure discomfort, including difficulty swallowing after cervical discography and lingering back pain after lumbar discography. One should also inform the patient that there will probably be a flare-up in usual symptoms that lasts 2–7 days, but infrequently can last for an indefinite period of time. The patient should be instructed to call if he or she experiences fever, chills, or severe (or delayed) onset of pain. Patients are contacted by telephone 48–96 hours after discography to screen for any possible complications or side effects.

POTENTIAL RISKS AND COMPLICATIONS

CLINICAL TRIALS: LITERATURE REVIEW

Discography has been proposed as a potential solution to the diagnostic dilemma concerning which patients to treat surgically and at what segmental level. Lumbar discography for the diagnosis of abnormalities involving intervertebral discs has been used extensively as a diagnostic tool for evaluating low back pain (LBP) since the 1950s. Although new diagnostic imaging tools have been developed and are widely used, discography is still practiced. Discography remains particularly useful in problematic cases unresolved by MRI or myelography, and for patients for whom surgery is contemplated.⁶ In 1992, Osti and Fraser³⁹ compared MRI with discography in disc disease and found that, using the current standard techniques, MRI failed to demonstrate some structural changes in the anulus that were visualized by discography. A similar result was obtained in a comparison of high-resolution CTs with discography.⁴⁰ Several studies have demonstrated that MRI cannot reliably predict which discs are painful upon discography, at least not to the level of confidence required to rely solely on MRI for surgical decision making.^41,42 A high intensity zone (HIZ) in the posterior anulus visualized by MRI has been proposed as a marker for painful discs.^11,43 While highly specific, the sensitivity of this finding is only 26%, limiting the usefulness of the HIZ in selecting patients for surgery.⁴⁴

While numerous papers have examined the usefulness of discography, some physicians still question its reliability.^45,46 Critics point towards mismatches between morphological features and clinical complaints and false-positive rates. Discography, especially combined with CT scanning, is exquisitely sensitive for identifying morphologic disc abnormalities, and the frequency of discography-detected abnormal discs in the LBP population is quite high. For example, Grubb et al.²⁹ found that in 78% of their patients with LBP, discography afforded positive findings for abnormality at one or more levels, whereas plain radiography and myelography demonstrated no such disease. Brodsky and Binder²³ also found that discography revealed abnormalities in many patients with normal-appearing myelograms. The frequency of morphologic abnormalities revealed via discography in the back pain population is high, and increases with age,^47,48 putatively due to painless degenerative changes.⁴⁸ In addition, the location of an annular tear and the location of pain appear to be uncorrelated.⁴⁹ Discrepancies between morphological appearance and pain provocation have also been described.⁵⁰ Millete and Melanson⁵¹ reported retrospectively that concordant pain was evoked by injection in only 37% of patients with a morphological abnormality documented by discography. Antti-Poika et al.⁵² reported only a 52.8% concordant pain provocation rate in discs where discography showed abnormal morphological changes. There have been some reports that only anulus tears may be reliably associated with pain provocation, and that other degeneration patterns were not necessarily associated with a pain response. It is clear that abnormal morphological structures revealed by discography are considered too non-specific to be clinically useful, and that positive results should be limited to those eliciting concordant pain.^46,53

Several series of asymptomatic patients have shown abnormal morphological findings using discography and CT–discography.^53–55 Sachs et al.¹⁰ reported a 13% incidence of abnormal disc structure detected by postdiscography CT scanning without pain provocation in a large series of patients. Several authors have also discussed that severe back pain may be elicited by discography in patients without a prior history of such pain.^45,54,55 Early work by Holt⁵⁵ reported a 36% rate of positive discography in asymptomatic subjects, although this study contained methodological flaws. These findings were subsequently refuted by Walsh et al.,⁵³ who demonstrated a 0% rate of positive discography in asymptomatic volunteers and established reproducible criteria for positive discography. Walsh studied 10 asymptomatic subjects and 7 patients with chronic LBP. Criteria for a positive result included 3/5 pain intensity (using a pain thermometer), similar distribution of pain, two types of pain behavior assessed by videotape review, and structural degeneration. Antti-Poika et al.⁵² reported in a prospective study of 279 injected discs in 100 patients that injection elicited concordant pain in >13% of patients with normal disc anatomy.

Caragee et al.⁵⁴ expanded on Walsh’s study of asymptomatic subjects by examining a cohort without LBP but with clinical characteristics closely matching those of patients with LBP. Results showed that, in individuals with normal psychometrics without chronic pain, the rate of false positives is very low if strict criteria are applied, and that false-positive rates increase with abnormal psychometrics and increased annular disruption.

The authors have recently studied the reliability of discogram using precise criteria for positive discography (e.g. reproducible NRS pain over 6/10 with <50 psi intradiscal pressure and <3.5 mL total volume) in 13 volunteers including normal asymptomatic participants. No false-positive results were recorded for examined discs. Pressure-controlled manometric discography coupled with precise criteria for positive discography can afford no false positives for the diagnosis of discogenic LBP in patients with normal psychometrics.

Surgical fusion has been used for pain relief in discogenic LBP unresponsive to conservative therapy. The identification of exact pain generators has been challenging, and many studies have investigated the reliability of discography for determining pain sources in LBP patients. Colhoun et al.⁵⁶ studied 195 patients with axial pain and reported that of 137 patients with a discogram positive for disc disease and provoked concordant pain, 89% derived significant, sustained clinical benefit from operation. Twenty-five patients showed morphologic disc abnormalities, but no provocation of concordant pain on discography. Among this group, only 52% had clinical success. Blumenthal et al.⁵⁷ reported that 74% of patients with internal disc disruption returned to work following anterior lumbar fusion performed based upon discography. In a multicenter surgical and nonsurgical outcome study after pressure-controlled discography, Derby et al.¹² stated that precise prospective categorization of positive discographic diagnoses may predict treatment outcomes, surgical or otherwise, thereby greatly facilitating therapeutic decision-making. In addition, patients with highly sensitive discs at low pressure appear to achieve significantly better long-term outcomes with interbody/combined fusion than with intertransverse fusion. Finally, although imperfect, discography is relatively safe, shows substantial sensitivity for identifying painful discs, and may predict surgery-related outcomes.

SUMMARY

Lumbar discography has been widely used for evaluating discogenic LBP. However the diagnostic power of discography remains controversial.⁵⁸ Walsh et al.⁵³ showed that the use of strict criteria for a positive test, including review of videotaped responses, affords excellent accuracy for the diagnosis of discogenic LBP. When excluding patients with somatization, applying more precise criteria for determining a positive level, and using a manometric grading scale, the specificity of lumbar discography increases 80–100%.³ Based on these data, Caragee et al.⁴⁵ concluded that in individuals with normal psychometrics without chronic pain, false-positive rates increase with abnormal psychometrics and increased annular disruption, and that the rate of false positives is very low if strict criteria are applied. False-positive rates of 40–50% in discogram studies of patients without a history of LBP have been obtained.^54,59,60

Since these studies, there have been advances in discography techniques and diagnostic criteria. One important recent advance has been the introduction of pressure-controlled manometric discography, which permits control of intradiscal pressure and the rate of contrast dye injection. Previous discogram studies assessing false-positive rates have limitations which include use of manual injections, uncontrolled injection rates, unrecorded and/or unreported opening and dynamic pressures, volumes, and maximal volumes. The authors have examined the diagnostic power of lumbar discography in select participants without present low back symptoms using advanced discogram techniques and criteria (R Derby et al., unpublished work, 2004). Results have shown that when using pressure-controlled techniques and precise criteria for positive discography (e.g., reproducible VAS pain > 6/10 with <50 psi intradiscal pressure, and <3.5 mL total volume) no false positives may be obtained for the diagnosis of discogenic LBP. Accordingly, discographers should carefully scrutinize positive responses obtained without pressure-controlled manometric discography and precise criteria. Strict attention and adherence to operational criteria reduces the risk of false-positive responses.¹³ The authors recommend positive criteria for lumbar pressure-controlled manometric discography as follows: numeric rating scale of pain above 6/10, <50 psi intradiscal pressure above opening pressure, <3.5 mL total volume, and at least one negative control disc.

Further, cervical discography is a safe and valuable diagnostic procedure if done by the highly trained and competent interventionist with expertise in the procedure.¹⁴ It is not a screening procedure but rather a confirmatory one. Even MRI is a useful adjunct to cervical discography; there are some MRI patterns that cannot be considered pathologic, and discography is required to diagnose discogenic pain syndrome.⁶¹ In a ‘select group’ of patients with chronic intractable neck pain but negative or indeterminate imaging findings who are being considered for surgical intervention, it can help localize the symptomatic level and potentially benefit the patients by surgical intervention.⁶² Finally, thoracic discography can be performed safely by experienced interventionists as a reliable tertiary diagnostic procedure to determine if degenerated discs on MR studies are related to clinical complaints.¹⁹ It can reveal the true source of pain and thus lead to precise and effective treatment.⁶³