SPINE Volume 21. Number 3, pp 300-312 ©1996, Lippincorr-Raven Publishers
Cervical Discogenic Pain Prospective Correlation of Magnetic Resonance Imaging and Discography in Asymptomatic Subjects and Pain Sufferers Kurt P. Schellhas, MD,* Michael D. Smith, MD,t Cooper R. Gundry, MD,* and Steven R. Pollei, MD*
Study Design. Asymptomatic subjects and chronic head/neck pain sufferers were studied with high-field magnetic resonance imaging and cervical discography to compare and correlate both tests. Objectives. To assess the accuracy of magnetic resonance imaging and discography in identifying the source(s) of cervical discogenic pain. Summary of Background Data. Previous retrospective studies describe a generally poor correlation between magnetic resonance imaging and provocative discography in the cervical spine. Methods. Ten lifelong asymptomatic subjects and 10 nonlitigious chronic neck/head pain patients underwent discography at C3-C4 through C6-C7 after magnetic resonance imaging. Disc morphology and provoked responses were recorded at each level studied, Results. Of 20 normal discs by magnetic resonance from the asymptomatic volunteers, 17 proved to have painless anular tears discographically. The average response per disc (N = 40) for this group was 2.42, compared to 5.2 (N = 40} for the neck pain group. In the pain patients, 11 discs appeared normal at magnetic resonance imaging, whereas 10 of these proved to have anular tears discographically. Two of these 10 proved concordantly painful with intensity ratings of at least 7/10. Discographically normal discs (N - 8) were never painful (both groups], whereas intensely painful discs all exhibited tears of both the inner and outer aspects of theanulus. Conclusions. Significant cervical disc anular tears often escape magnetic resonance imaging detection, and magnetic resonance imaging cannot reliably identify the source(s) of cervical discogenic pain. [Key words: discography, magnetic resonance imaging, cervical spine, cervical pain, disc degeneration] Spine 1996;21:300-312
Magnetic resonance (MR) imaging has become widely accepted as a valuable imaging study in the investigation From the ""Center for Diagnostic Imaging, St. Louis Park, and the f Minnesota Spine Center, Minneapolis, Minnesota. Supported in part by grants from the Minnesota Spine Center Research Fund and the Center for Diagnostic Imaging. Acknowledgment date: February 27, 1995. First revision date: April 27, 1995. Acceptance date: June 23, 1995. Device status category: 1.
300
of spinal disorders.1'13'14 Discography has been and continues to be used by clinical practitioners in selected cases as both a provocative and morphologic test to evaluate suspected discogenic pam_3-6~s'1:M2,i6-26 "The anatomic basis of primary cervical discogenic pain provoked with discography has been described.2'5 Prior investigations in the thoracic17 and lumbar spine21*23 have proven that, when performed by a skilled and experienced discographer, discography is safe and accurate in distinguishing painful, symptomatic, and anatomically deranged discs from asymptomatic discs, be they either normal or deranged. Preceding investigations describe a high incidence of cervical disc lesions revealed discographically with variable provocative responses and MR appearances.12'16 To compare and correlate MR imaging and discography in the evaluation of cervical or head pain, we studied 10 lifelong asymptomatic adult volunteers with both modalities and compared these results to those from a similar group of adult, nonlitigious chronic head and neck pain sufferers. • Materials and Methods A paid advertisement was placed in a regional university newspaper (Minnesota Daily, University of Minnesota, Minneapolis, MN) recruiting adult subjects at least 21 years of age, willing to undergo an'invasive test involving puncture and injection of cervical discs. Prerequisites for participation in the study were: 1) a lifelong negative medical history pertaining to either chronic or recurring headaches, neck, shoulder, upper back, or upper extremity pain, upper extremity numbness, paresthesla, restricted neck morion, or history of prior neck injury; and 2) lack of any complicating medical problems, such as bleeding disorders, history of mental illness, drug abuse, or active Infectious diseases. Patients were initially offered S250.00 cash to be paid on completion of the study. Over 70 subjects responded to the advertisement and were further interviewed by telephone regarding medical history, general body habitus, height, and weight, and the study itself was explained in more depth to them. We attempted to select slender people who could easily extend their neck and thus be ideal candidates for undergoing cervical discography. We avoided unusually large people or subjects with short, thick necks who might pose a technical problem in completing the multilevel examination.
Cervical Discography and MR * Schellhas el al After initial screening by telephone, eligible subjects were physically examined and interviewed by an experienced spine surgeon. Subjects who passed rhis screening interview subsequently underwent a lateral cervical radiograph, were asked to sign the consent form to participate in the study, and were then scheduled to go to the outpatient Imaging center to undergo a screening MR imaging study of the cervical spine, to be followed immediately (the same day) by multilevel cervical discography .involving the C3-C4 through C6-C7 discs. They were informed that the study would involve videotaping their face during injection of each cervical disc. Ten subjects (six women, four men, aged 21-41 years, with a mean age of 30.4 years) were selected for the formal investigation. All subjects arriving for the actual cervical discography were initially interviewed by the discographer about their medical history to see if anything had changed since the initial interviews. Subjects thereafter underwent high-field MR im-'. aging using a 1.5 T superconducting magnet (General Electric, Milwaukee, Wl), employing sagittal, 3-mm thick, 451/14 (TR/ TE) multiplanar gradient refocused images of the cervical spine. We elected to perform sagittal studies only because of cost considerations and our experience that axial images seldom reveal disc anular lesions not visible on sagittal views. The sagittal sequence that we used has been routinely used in our daily practice for over 5 years, and we believe it to be the most sensitive sequence for detecting cervical disc lesions. The previously obtained lateral cervical spine radiograph and the sagittal MR images were then reviewed by the procedural neuroradiologist discographer for possible discography contraindications, such as spina! cord compression, after which subjects were taken into the special procedure suite. The procedure of discography was again explained to them in depth, and an additional consent for the discography procedure was signed. Discography was performed by a single, highly experienced, board-certified, procedural neuroradiologist, righthanded discographer. Examinations were performed in a cus'tom-designed procedural suite with a high-resolution, multidirectional C-arm fluoroscopic apparatus with 4.5-, 6-3 and 9-inch image intensifier modes. The patient's face was videotaped during each of the cervical disc injections. Discs were studied in random fashion with the subjects unaware of which disc was being studied. The video camera operator was given printed indication of which disc was about to be injected, and this was recorded on tape before the actual disc injection. Subjects were placed on the fluoroscopy table supine, with their shoulders slightly elevated to allow extension of the neck. The subject's chin was turned approximately 15° to 30° toward the left, opposite the discographer, and a piece of tape was placed from the table across his or her forehead to prevent any sudden movements. The discograms were all performed from the patient's right-hand side. After subject positioning, the skin on the neck was marked using a ballpoint pen under fluoroscopy. The C6-C7, C5-C6, C4-C5, and C3-C4 discs were identified under fluoroscopy, and the skin overlying the most ideal access to each disc was indented with the back end of the ballpoint pen, leaving a rounded, target-like indentation in the skin. After skin marking, the subject's skin was then thoroughly cleansed twice with iodine solution, followed by generic isopropyl alcohol rinse (X2), after which a sterile fenestrated stick-on drape was applied.
301
Immediately before needle insertion, the patient's skin was again rinsed with alcohol. Needle placement into each disc was performed by initially palpating the subject's neck with the second and third digits of the left hand, while the 25-gauge spinal needle was held with the right hand. Subjects were often asked to gently swallow during palpation to facilitate carotid artery localization and movement. The carotid artery was gently pulled laterally in most cases, while the esophagus was pushed medially, allowing the spine to be directly palpated. The 25-gauge, 3-inch spinal needle (Becton Dickinson, Franklin Lakes, NJ) was carefully advanced either into the disc or to the cervical spine near the disc, passing the needle through the skin mark that had been placed earlier. In most cases, the disc was not punctured on the initial needle placement. If the disc was not entered initially, a fluoroscopic check was performed to assess needle position along the anterior margin of the spine. If the needle tip was noted to be near the disc in question, attempts were made to manipulate the needle incrementally into the disc by rotating the bevel and making small withdrawal and advancement movements. This was not performed under direct fluoroscopy to avoid exposure of the discographer's hands. After successful disc puncture, needle position was then assessed with fluoroscopy, and the needle tip was advanced into ideal position, as near the center of the disc nucleus as was possible. In most instances, needle placement into each disc took on the order of 10 to 60 seconds. After optimal needle positioning, preparation was made for disc injection. The camera operator was alerted to begin videotaping, and was notified as to which disc had been punctured with a piece of paper on which was written the disc identity, out of the subject's view. The needle stylet was then removed and contrast (iohexol; Nycomed Pharmaceuticals, New York, NY), 180 mg/ml in a 5-ml syringe using a 40-cm connecting tube, was attached to the needle. The subject was then instructed to discount any sensations that he or she had felt during placement of the needle and focus exclusively on the upcoming injection. Injection was then performed under simultaneous live fluoroscopic observation of the disc and videotaping of the subject's face. Fluoroscopy was performed either in an anteroposterior or lateral projection, depending on what fluoroscopic projection had been last used for checking needle position within the disc. Discs were injected either to capacity, until leakage (venous, epidural, or paraspinous) of contrast was seen, or until an obvious manifestation of pain was observed on the subject's face (none in this entire group). Injection volume, presence or absence of injection endpoint, and firm versus soft endpoint characteristics were all recorded. The presence or absence of an externally visible facial response (e.g., grimace, wince) was not initially recorded; however, the tape was later reviewed for this information. Anteroposterior and lateral magnified fluoroscopic spot films of each disc were obtained, and if injection was clearly intranuclear and films satisfactory, the needle was then removed. After needle removal, subjects were immediately interviewed in depth about the experience during injection. Subjects were asked to describe In detail the exact type of sensation (e.g., pain, pressure) and its location (e.g., head, neck, shoulder, chest, upper extremities, front, back) during disc injection, excluding needle placement and removal. They were asked to rate the intensity of this experience on a scale of 0 (no sensation whatsoever) to 10 (extreme pain, pressure, and the like), and describe whether this was similar (concordant) or dissimilar (nonconcordant) to
302
Spine • Volume 21 • Number 3 • 1996
Rgure 1. Completely normal magnetic resonance and discography (see Rgure 4) in 28-year-old lifelong asymptomatic volunteer. Sagittal, 3-mm thick, multiplanar gradient refocused midline image reveals normal discs, spinal cord, and centra! spinal canal dimensions.
their clinical complaints (none for this group). This intensity was recorded, after which the subject's skin was again rinsed with alcohol and discography was performed at subsequent levels, using the same technique. Five selected patients from the study underwent high-resolution, thin-section computed tomography (CT) of selected injected discs, in which unsuspected anatomic abnormalities were observed. Subjects were carefully examined by the discographer for any evidence of subcutaneous bleeding or hematoma formation immediately after and again 20 and 40 minutes after discography to make sure diat there were no complications. After the second postprocedure check, subjects were advised of any possible side effects to watch for during the next several days, and were allowed to leave the department. All subjects were offered a nonrenewable prescription for 20 narcotic analgesic tablets, to be taken as needed for possible discomfort. They were thereafter contacted by telephone 48 to 96 hours after discography to screen, for any possible complications or side effects. Ten consecutive, nonlltigious, adult (aged 21—48 years, mean age 37.8 years, six women, four men) chronic neck and head pain sufferers also underwent four-level discography involving the C3—C4 through C6—C7 discs, using the identical discography protocol for positioning, skin cleansing, needle placement, injection, postinjection interview, and filming. All
10 of these subjects had suffered pain and symptomatology for a minimum of 6 months before referral for discography. These patients also had all undergone high-field MR imaging of the cervical spine within 90 days (four on the same day) of discography without change in clinical status during the interval. In this group, the first five examinations were videotaped. In the five subsequent cases, which were not videotaped, note was specifically made of any externally visible pain responses during injection, and the tapes of the earlier patients were reviewed for this specific information (Tables 1 and 2). After completion of each discogram, the MR and discography examinations were interpreted by the radiologistdiscographer. For purposes of formal investigation, the MR and discography studies from both groups were subsequendy interpreted blindly and independently by two board-certified and highly experienced spine radiologists, other than the discographet, both completely blinded to patient and group identity, disc injection characteristics, and the provocative discography results. Each formal interpretet initially read the MR studies, after which they submitted their written results before interpreting the discographic morphology. MR disc morphology was categorized as either normal or abnormal. Discographic morphology was categorized according to a much-simplified modification of the Dallas Cervical Disc Morphology Classification.11 Categories included normal morphology, localized inner anular fissuring, and major tears involving both the inner and outer aspects of the anulus, either localized or extensive to circumferential, with or without obvious leakage of contrast into either the epidural space, veins, or paraspinous tissues. A simplified discographic morphology classification (three categories) was used because only selected discs underwent postdiscography CT owing to cost considerations, and some of the intensely painful discs were not injected to capacity, definitely influencing the morphologic appearance on radiographic spot films in some instances. After both radiologists-had submitted their written interpretations, they met and came to mutual agreement at individual levels where their initial interpretations had differed. Only the final agreed-on interpretations were submitted as official. •
Results Asymptomatic Volunteers
Ten lifelong asymptomatic subjects (six women, four men, aged 21—41 years, mean age 30.4 years) underwent high-field MR imaging of the cervical spine, followed immediately thereafter by videotaped discography at C3—C4 through C6— O7- There were no complications in this series. None of the lifelong asymptomatic subjects experienced any iatrogenic pain or discomfort later than the evening of the procedure. An analysis of the MR morphology, discographic morphology, and provocative responses to injection at each disc level is presented in Table 1. Initial interobserver agreements with regard to MR and discographic morphology at each disc level were 88.75% and 91.25%, respectively. After both formal interpreters had met and mutually agreed on any initial interpretive discrepancies, MR examinations were judged to be normal in only 1 of the 10 patients (Figure 1), whereas 9 of the 10 MR studies exhibited at least one abnormal disc (Figures 2
Cervical Discography and MR • Schellhas et ai
303
Table 1. Lifelong Asymptomatic Volunteers
Patient Age (yr)
Sex
28
F
41
22
M
M
36
M
28
F
31
36
M
F
26
F
21
F
35
F
•
Disc Level
MR Morphology
Discographic Morphology
C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C&-C7
0 0 0
1 ! 0 0 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 1 2 2 2 0 2 2 2 0 2 2 2 1 2 2 2 2 2 2 0 2
0 1 0 1 0
1
0
1
1 1 1 1 0 0 0
1
1 0
1
0
1 1 0
1 0 0
1 1
0 0
1 0 0
1 1 1
1
Intensity of Response to Injection
2 2 4 4 3 4
2 0 0 6 2 2 2 2 4 2 6 1 3 1 2 3 2 0 5 3 2 1 4 2 0 0 3 4 1 1 2 2 4 3
Nonconcordant/ Concordant
Facia! Expression
0
0 0 0 D D 0 0 0 0 0 0 0 0 0 0 0 0 D 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
0 0 0 0 0 0 D 0 D 0 0 0 0 0
D D D D D 1 0 0 0 D D 0 D D 0 0 0
MR morphology: 0 = normal; 1 - abnormal. Discographic morphology: 0 = normal; 1 = inner anular tear: 2 = inner and outer anuiar tear. Intensity of response to injection; 0 - no perception of pain/pressure on 0-10 scale; 10 = severe pain/pressure. NonconcordanVconcordant: 0 = nonconcordant; 1 = concordant. Facial expression: 0 = none; 1 = externally visible manifestation of pain. MR = magnetic resonance.
and 3). There were no frank cervical disc herniations in this group. The mosr commonly observed disc abnormalities on MR imaging were bulging anuli (Figures 2, 3). Twenty discs from this group were judged to be morphologically normal on MR, whereas 20 discs exhibited morphologic abnormality. Discographic contrast injection volumes ranged from 0.2 to 0.4 ml with firm endpoint for normal discs, to more than 1.2 ml for leaking discs. Analysis of discographic morphology revealed only five completely normal discs (Figure 4), indicating that internal disc derangement and various degrees of anular disruption are common in cervical discs of asymptomatic subjects (Figures 5—7). The most common discographic abnormality observed was disruption of the inner and outer aspects
of the disc anulus posterolaterally (Figure 6), with or without full-thickness disruption and leakage of contrast into the extraanular spaces (Figures 5, 7). Extraanular leakage of contrast was noted in 17 of 35 morphologically abnormal discs. Despite this high prevalence of anatomic disc derangement, there were no intensely painful discs in this entire series. The highest pain intensity rating was 6/10, and was entirely unfamiliar to the patient. The average intensity of perception to injection per disc from this entire group (N = 40) was 2.42 (SD = 1.5). There were no externally visible facial manifestations of pain during injections in this series (0/40). The most frequently encountered intensity of perception/response to injection was 2 to 2.5 (N = 14). Five disc injections failed to produce any sensation (0/10).
304
Spine • Volume 21 - Number 3 • 1996
Table 2. Pain Sufferers Patient Age (yr)
Sex
33
M
21
F
40
M
28
40
48
42
39
46
35
F
F
F
M
M
F
F
Disc Level
MR Morphology
Discographic Morphology
C3-C4 C4-C5 C&-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 C5-CG C5-C7 C3-C4 C4-C5 C&-C6 C6-C7 C3-C4 C4-C5 C5-CS C5-C7 C3-C4 C4-C5 C5-C6 C6-C7 C3-C4 C4-C5 CS-C6 C6-C7 C3-C4 C4-C5 C5-C5 C6-C7
0
2 2 2 2 0 1 1
0 1 1 1 0 1 D 1 1 1 0 0 0
] 1 0
1 1
1
1
1 1
1 0 0
1 1
1 1 1 1
1
1
1 1 1 1 1 0
0 2 2
2 2 2 2 2 0 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
Intensity of Response to injection
3 2 0 5 5 2 7 0 10 S 6 7 5 9 8 2 5 8 10 8 0 3 2 2 5 5 10 8 2 3 8 7 6 9 3 2 4 8 ID 5
Nonconcordant/ Concordant
Facial Expression
1 0 0 1 0
1
0 0 0 0 0 0 1 0 1 0 0
0
0
1
1 1
1 1 0
1 1 1
1
1
0
0
1
0
1
1 1
0
0
1 1 1
0
0 0 0
1 1
0 0
1 1 0
1
1
0 0 0 0 0
1
1 0 0
1
1 1 1
0 0 0
0 0 0
1 1 1
1 1 0
MR morphology: 0 = normal; 1 = abnormal. Dlscographic morphology: 0 = normal; 1 = inner anular tear; 2 = inner and outer anular tear. Intensity of response to injection: 0 = no perception of pain/pressure on 0-10 scale; 10 = severe pain/pressure. Nonconcordant/concordant: 0 = nonconcordant; 1 = concordant. facial expression: 0 = none; 1 = externally visible manifestation of pain. MR = magnetic resonance.
One clinically familiar (concordant) sensation of trapezius muscle "cramping11 was provoked during injection of a deranged C3—C4 disc exhibiting an ipsilateral inner and outer anular tear (Figure 5A, B). This patient stated that we had not specifically asked her about muscle "cramps" in the neck and she had entirely forgotten about ever having had such an experience until the disc was actually injected. She denied any pain with injection, and reported only a level 4/10 inrensity sensation of cramping within the muscle. Clinical Pain Sufferers
Ten nonlitigious, chronic head and neck pain sufferers (six women, four men, aged 21—48 years, mean age 37.8 years) with symptoms of at least 6 months' duration also underwent four-level cervical discography
wirhin 90 days of MR (four patients on the same day). None of these patients had experienced any change in symptomatology between MR imaging and discography. MR and discographic findings for this group are summarized in Table 2. Eleven discs in this group were interpreted as normal on MR, whereas the remaining 29 exhibited variable abnormality (Figures 8, 9). Of the 11 discs that were normal on MR imaging, 10 of these proved to have anular tears discographically. All discograms on this group were performed without complication. Responses to injection were considerably different for this group compared to the asymptomatic subjects. Intensity of response to injection was arbitrarily divided into three caregories: response of 0 to 3 was minimal, 4 to 6 was moderate, and 7 to 10 was intense.
Cervical Discography and MR • Schellhas et al
305
nally visible pain, and hence the disc was not fully distended, suggesting the possible existence of a more extensive tear. An important observation pertaining to provocation was that both axial and radicular pain was observed. We frequently provoked extraspinal pain in radicular distribution, as well as distant or referred pain such as in the mastoid, mandible, temporomandibular joint, occiput, and anterior chest (Table 3). Most often, radicular or referred pain was observed with axial (neck) pain; however, pure extraaxial pain was also provoked with injection. These perceptions were always associated with anuiar tears, usually toward the side of perception. •
Figure 2. Multiple bulging discs in asymptomatic 35-year-old female volunteer. Adjacent (A on left, B on right), sagittal, 3-mm thick1 images reveal perceptible anuiar bulging at C3-C4 through C6-C7, as formally interpreted (Table 1). Painless tears found discographically at C3-C4, C4-C5, and C6-C7. C5-C6 discographically normal.
Another clear difference between this symptomatic group and the volunteers was that disc injection was voluntarily terminated before either reaching an endpoint or achieving optimal opadfication in 12 discs because of obvious manifestations of pain or distress due to injection. There were no such voluntary injection terminations due to pain or distress in the asymptomatic group. In selected instances, lidocaine 2%, 0.3 to 1.0 ml, was injected into especially painful discs after filming, before needle removal.. Eight of the 10 anuiar tears found at discography in MR normal-appearing discs proved concordantly painful when injected. Two of these deranged, concordantly painful discs that were normal on MR had intensity ratings of 7/10 or greater. Fifteen discs from this group exhibited a combination of intense, concordant pain and anuiar tears at discography. There was external manifestation of pain in all 15 of the intensely painful discs (Table 2). Thirteen of these 15 intensely painful, concordant discs were interpreted as abnormal on MR imaging. Sixteen clearly MRabnormal discs from this group proved to be neither painful nor concordant at discography, with two exhibiting normal morphology (false-positive MR). Discographically normal morphology discs (N = 3) were never painful, whereas 14 of 15 intensely painful discs exhibited tears involving both the inner and outer aspect(s) of the disc anulus, most with leakage of contrast outside of the disc (Figures 10, 11}. A single, intensely painful, concordant disc from this group was blindly interpreted to exhibit only an inner anuiar tear (Table 2). The injection of this disc had been voluntarily terminated because of the immediate provocation of exter-
Discussion
Attempts to identify a characteristic marker of cervical discogenic pain with MR imaging have been unsuccessful to date,12'16 although it is suggested that cervical discs that appear abnormal on MR imaging are more likely to be painful than those that appear normal. These cervical observations differ from the lumbar spine, where the lumbar disc high-intensity zone1'18 has been identified as a highly reliable marker of discogenic pain in symptomatic subjects. In addition to the lack of any characteristic MR disc finding, highly variable provocation responses have been described in the cervical spine, with disc fissures and full-thickness anuiar tears associated with contrast leakage.16 With these variables in mind, discography remains an important tertiary procedure for investigating suspected cervical discogenic
Figure 3. Multilevel cervical disc degeneration with painless anular tears at discography {see Figure 5) in 28-year-old asymptomatic volunteer. Adjacent, sagittal, 3-mm thick multiplanar gradient refocused images through the midline (A) and to left of midline (B) reveal small disc protrusions at C5-C6 and at C6-C7 (straight arrows}, with minimal bulging of the C3-C4 anulus an the left (curved arrow in B).
306 Spine • Volume 21 • Number 3 • 1996
Rgure 4. Normal discography in 28-year-old volunteer (same patient as Rgure 1). Normal anteroposterior (A, C) and lateral (B, D] discography films. Note the tight rounded configuration of the nucleus in these normal discs. Nonconcordant, low-intensity pressure was perceived during discography at each level in this subject
pain. MR imaging is a valuable screening test, but our study proves that it is not definitive when assessing patients whose primary complaint is axial or radicular pain without neurologic deficit. In cases where surgical intervention is being considered, the importance of preoperative identification of the exact sources of pain generation must be emphasized. In our daily practice, we often perform cervical discography on patients who have already undergone one or more cervical operations, which had been based on imaging studies alone, where the postoperative result was less than satisfactory. In most of these cases, we provoke the original pain at unoperated, often inconspicuous-appearing (on imaging studies) discs. This may imply that the previous unsuccessful operations could have been circumvented had discography been performed. Based on our clinical experience and the results of this study, which demonstrate the unreliability of MR at identification of painfully deranged discs, we may perform discography as an important adjunct in those patients with primary complaints of unrelenting pain without neurologic deficit before contemplated surgical
intervention. From our clinical experience only, we believe that the best operative candidates after cervical discography are nonsmoking, otherwise healthy patients with either single-level or adjacent concordantly painful and deranged discs with painless normal or minimally deranged neighboring discs (transitional levels after surgery). Cervical discography excludes contemplated surgery in most of the patients we study. Clearly, formal, prospective, long-term analysis of outcomes is warranted in this regard. Another situation in which discography may prove valuable is in chronic pain patients whose CT or MR imaging studies are either negative or equivocal and discogenic pain is suspected. The establishment of an accurate diagnosis can aid in patient counseling and treatment planning. In addition, the judicial practice of awarding or denying compensation for pain to claimants based on imaging studies alone appears flawed and unjustified in light of our findings. We observed a high incidence of painless cervical disc anular tears at discography in our carefully screened, lifelong, asymptomatic subjects (Figures 5—7), as well as in the control group of chronic pain sufferers. Despite
Cervical Discography and MR * Schellhas et al
307
data reveal that cervical-discograms in truly asymptomatic adult volunteers are not painful, regardless of the presence or absence of incidental disc derangement and anular disruption. We clearly differentiated the asymptomatic group from the clinical group of pain sufferers in terms of average response per disc injection, suggesting that in skilled and experienced hands, discography can be relied on to identify clinically significant cervical disc anular lesions, as it can be in the lumbar region.21*23 In our entire study (20 subjects, 80 discs), 27 of 31 discs that appeared normal on MR imaging (Figure 1) proved to have anular tears of varying degree (inner, inner and outer without leakage, and full-thickness with leakage), location, and extent around the disc anular circumference, indicating that MR in its current state of technology does not reliably detect anular defects. Most of these anular tears proved to be posterolateral or uncovertebral in location (Figure 6), as described in prior reports.11'12'16 In our control group of pain sufferers, 10 of 11 MR-normal discs proved to have sub-
Figure 5. Painless anular tears in asymptomatic volunteer (same patient as Figure 3). AntEroposterior (A) and lateral (B) discographic films revea! a partial-thickness foramina! anulartear on the right (large arrow in A), with full-thickness posterolateral tear on left {arrowheads in A and arrow in B). Injection.produced a a irapezius muscle "cramp" that was given a perception intensity rating of 4/10. Anteroposterior (C) and lateral (D) discographic films reveal a large epidural leak of contrast accumulation (arrows) through a full-thickness posterolateral anular tear. Note optimal needle positioning within the center of disc nucleus. Injection produced a 3/10 perception rating score of pressure.
the fact that anular tears are common in both groups and not necessarily associated with pain, we do observe a progressive incidence of concordant high-intensity axial and radicular pain at discography when we move from completely normal discs (Figure 4) to full-thickness posterior anuJar rears (Figures 10,11) in symptomatic patients. Inner and outer to full-thickness, leaking anular tears (Figures 5, 7) are both common and inconsequential in asymptomatic subjects, whereas tears of this type were demonstrated in 14 of 15 of the intensely painful and concordant discs in the pain sufferers. The locarion of perceived pain is variable, but usually on the same side as the obvious anular tear (Table 3), Conversely, completely normal discs are both uncommon at discography, regardless of the patient's clinical status and history, and they are not painful when injected. Our
Figure 6. Painless foraminal anulartears in asymptomatic subject. Antero posterior discography film reveals needle in C4-C5 disc with right-sided foraminal anulartear and contrast accumulation (upper arrow). Note similar foraminal tear at C5-C6 (lower arrow). Magnetic resonance imaging was normal at C5-C6 and minimally abnormal at C4-C5.
308 . Spine • Volume 21 • Number 3 - 1996
Rgure 7. A, B, Painless, circumferential, cervical anular disruption with leakage of contrast in 41-year-old male asymptomatic volunteer, with multilevel cervical disc degeneration. Magnetic resonance imaging revealed degeneration and anular bulging at this level. Note circumferential anuiar tear (arrows). Nuclear compartment grossly enlarged and degenerated (compare with Rgure 4). Injection produced nonconcordant pressure sensation rated at 3/10.
stantial anular tears at discography, with two of these intensely and concordantly painful. These results demonstrate that a normal MR imaging study of either the entire cervical spine or of any particular disc does not exclude the existence of clinically significant disc disease in patients suffering chronic neck, head, or radicular pain. Thin-section sagittal MR images13'14 provide the most information'regarding disc integrity and contour. In our experience, thin-section MPGR (as we used in this study) sagittal views are the most sensitive at de-
tecting intrinsic and anular disc lesions. We were unable to identify any of the unexpected, but discographically proven, anular tears in the pain sufferers when we reviewed axial MR images (not part of this formal investigation) after discography. We have usually found that postdiscography CT scans (Figure 10D) are noncontributory. This opinion is supported by others16 and suggests that postdiscography CT scanning should be abandoned as a routine procedure.
Rgure 8. Minimally abnormal magnetic resonance imaging study with multiple painful tears at discography performed the same day (see Rgure 10). Adjacent, sagittal, 3-mm thick multiplanar gradient refocused images through the cervical spine in the midline (A) and to left of midline (B) reveal normal-appearing discs with a large central spina! canal and normal spinal cord. Minimal (interpreters initially split) anular bulging (opposite the discographically identified tears) evident laterally on images not shown.
Figure 9. Mildly abnormal magnetic resonance imaging at C6-C7, but normal at C5-C6, where discography produced a 10/10 pain perception rating, which was concordant {see Figure 11). Adjacent, 3-mm thick multiplanar gradient refocused images through the cervical spine to the right of midline (A) and in midline (B) reveal slight vertical narrowing and circumferential anular bulging at C6-C7 (arrows). The higher cervical discs appear unremarkable. Spinal canal is of normal dimension.
Cervical Discography and MR * Schellhas et al
309
Figure 10. Painful anular tears not detected with magnetic resonance performed the same day (same patient as Figure 8). Anteroposterior (A) and lateral (B) films of the C5-C6 disc reveal leakage of contrast (arrows) into the epidural space, primarily to the left of midline (A). Severe concordant neck pain was provoked with injection. Anteroposterior discography film (C) reveals left-sided dorsal epidural leakage of contrast (arrows), similar to that observed at C5-C6 (A, B}. An axial, 3-mm thick postdiscography computed tomography scan (D) demonstrates the posterior and left-sided full-thickness anular tear (arrows) in the C6-C7 disc. The site of this anular tear is included in the view on Figure 8B. Severe, concordant, left-sided pain was provoked during discography.
Cervical discography is highly operator dependent, and our results reflect the experience and technical skill of the discographer. All of our discograms were performed by a single, highly experienced procedural neuroradioiogist who, before undertaking this study, had performed multilevel cervical discography on over 200 clinical subjects without any complications. These results differ from prior investigations,1 where a relatively low, yet reportable incidence of significant infectious complications has been described. Infectious risks10'15 exist with discography in general; however,
Figure 11. Concordantly painful (10/10) discography with fullthickness, leaking anular tear at C5-C6 (same patient as Figure 9). Anteroposterior (A) and lateral (B) discography films reveal optimal needle placement within the C5-C6 disc with a full-thickness posterior anular tear, allowing epidural leakage of contrast (arrow in B). C6-C7 disc also is degenerated (lower opacified disc in A). C6-C7 disc painfully concordant but of lower intensity than the C5-C6 disc.
these can be minimized by careful patient and disc selection and by adhering to the sterile technique we describe. We have never used routine antibiotics before, during, or after lumbar, thoracic,17 or cervical discography, as others have advocated.9 We believe that using a single 25-gauge needle for each disc is advisable for this procedure, in the unlikely event that a carotid, vertebral, or smaller artery is accidentally punctured. Patients should not have any bleeding problems if an artery is accidentally punctured by a needle of this caliber. It is imperative for the needle tip
310 Spine • Volume 21 • Number 3 * 1996 Tabie 3. Sites of Reported Pain/Pressure Perceptions Cervical Disc C3-C4
C4-CS
C5-C6
C6-C7
Sites Mastoid, temple, jaw, TMJ, parietal, occiput, craniovertebral Junction, neck, throat upper back, trapezius muscle, top of shoulder, upper extremity Mastoid, TMJ, parietal, occiput, craniovertebral junction, neck, throat, trapezius muscle, shoulder, upper extremity, anterior chest, upper back Occiput, craniovertebral junction, neck, upper back, throat trapezius muscle, shoulder, upper extremity, anterior chest Neck, upper back, scapula, shoulder, trapezius muscle, upper extremity, anterior chest
to be placed as near to the center of the disc as is possible (Figures 4—7, IDA—C3 11) to ensure a valid intranuclear injection. Extranuclear injections usually produce an unreliable response in our experience. Needle position was optimal and intranuclear in all 80 of the discs in our series. The issue of cervical discography contraindications merits discussion. We avoid performing both cervical and thoracic discography at disc levels producing spinal cord compression17 (Figure 12) with or without clinical myelopathy. It is our fear that injecting and distending a disc already compressing the spinal cord could potentially either produce or aggravate myelopathy. It is imperative that the clinician have knowledge of the central spinal canal dimensions before undertaking either cervical or thoracic discography to avoid potential neurologic complications. Ideally, the clinician should have recent (within 6 months and no change in clinical status) CT or MR imaging studies of the cervical (or thoracic) spine before performing discography. Patients with known bleeding abnormalities or those on anticoagulant therapy must be handled with special caution. In our practice, patients are always thoroughly questioned at the time of scheduling about their coagulation status and whether they are on any type of anticoagulant medication. We specifically ask about easy bruising, nosebleeds, slow stoppage of bleeding from cuts, and the like. In patients who are on anticoagulants, we contact their physicians and obtain permission to stop the medication 1 or 2 days before the discogram, and obtain laboratory proof of acceptable coagulation parameters before undertaking the procedure. Contrast discography followed by intradiscal analgesic injection at painful levels to abolish the pain has been advocated4 to define conclusively the relationship of disc lesion to clinical complaints. We perform analgesic disc injections during discography whenever possible to relieve prolonged discomfort. We do not use these analgesic injections for diagnostic purposes. We did not include analgesic injections in this formal study because of the fact that many painful discs could not be injected with analgesic after having been distended to capacity
with contrast. To include routine intradiscal analgesic injections in this study would have necessitated return visits and separate therapeutic procedures. Magnetic resonance imaging of the cervical spine cannot be relied on conclusively either to identify painful cervical discs or rule out anular lesions, and should be used only as a screening test in patients with primary complaints of neck, head, or radicular pain. Cervical discography represents a definitive test that can be performed reliably and safely by experienced individuals. If cervical discography is performed, we recommend studying all accessible discs from C3-C4 through C6— C7. If C2—C3 and C7-T1 can be safely studied, these might also prove valuable, depending on the results at
Figure 12. Cervical disc degeneration leading to spinal cord compression. Sagittal, 3-mm thick, midline multiplanar gradient refocused image reveals advanced disc degeneration at C4-C5 and C5-C6 with frank cord compression at C5-C6. Clinical myelopathy was present with extremity weakness and altered reflexes. It is our recommendation that cervical discography not be performed at levels where spinal cord compression exists, particularly if myelopathy is present. We successfully performed discography at C3-C4 and C6-C7 in this case.
Cervical Discography and MR • ScheJJbas et a] other levels. Clinical indications for cervical discography include cases of chronic neck, head, or radicular pain where normal, equivocal, or contradictory (lateralizing.pain complaints with contralateral imaging findings) imaging studies exist, provided that there is no procedural contraindication. References 1. April! C, Bogduk N. High-intensity zone: A diagnostic sign of painful lumbar disc on magnetic resonance imaging. Br J Radio! 1992;65:361-9. 2. Aprill C, Bogduk N. On die nature of neck pain. Pain 1993 ;54:213-7. 3. Aprill CN. Diagnostic disc injection. In: FrymoyerJW, ed. The Adult Spine: Principles and Practice. New York: Raven Press, 1991:403-19. 4. Bogduk N. Point of View. Spine 1994;19:2S24-5. 5. Bogduk N, Windsor M, Inglis A. Th'e innervation of cervical intervenebral discs. Spine 1988;13:2—8. 6. Brightbill TC, Pile N, Eichelberger RP, "Whitman M. Normal magnetic resonance imaging and abnormal discography in lumbar disc disruption. Spine 1994;19:1075-7. 7. Cloward RB. Cervical discography: Technique, indications and use in diagnosis of ruptured cervical discs. AJR Am J Radiol 1958;79:690-705. 8. Colhoun E, McCall IW, Williams L, Pullicino VNC. Provocation discography as a guide to planning operations on the spine. J Bone Joint Surg [Br] 198S;70:267-71. 9. Fraser RD, Osti OL, Yemen-Roberts B. latrogenic discitis: The role of intravenous antibiotics in prevention and treatment—an experimental study. Spine 1989;14:1025~32. 10. Guyer RD, Collier R, Sn'th WJ, et al. Discitis after discography. Spine 19SS;13:1352-4. 11. Handal JA, Holliday JG, Zardus M. Morphologic classification of cervical discograms. Poster Exhibit Award, North American Spine Sociery Meeting, Keystone, Colorado, Jul 31— Aug 3, 1991. 12. Handal J, Schleusner R,.Zardus M, Holliday G. Clinical correlation of cervical magnetic resonance imaging and cervical discography. Orthop Trans 1991;15:223-». 13. Modic MT, Masaryk TJ, Ross JS, Carter JR. Imaging of degenerative disk disease. Radiology 19S8;16S:177-86. 14. Modic MT, Weinstein MA, Pavlicek W. Magnetic resonance imaging of the cervical spine. AJR Am J Radiol 1983; 141:1129-36.
311
15. Osti OL, Fraser RD, Vernon-Roberts B. Discitis after discography. J Bone Joint Surg [Br] 1990:72:271—4. 16. Parfenchuck TA, Janssen ME. A correlation of cervical magnetic resonance imaging and discography/computed romographic discograms. Spine 1994;19:2819-25. 17. Schellhas KP, Pollei SR, Dorwart RH. Thoracic discography: A safe and reliable technique. Spine_l994;19:2l03—9. IS. Schellhas KP, Pollei SR, Gundry CR, Heithoff KB. Lumbar disc high intensity zone: Correlation of MR and discography. Spine 1996;21:79-86. 19. Shinomiya K, Nakao N, Shindoh K, Mochida K, Furuya K. Evaluation of cervical discography in pain origin and provocation. J Spinal Disord 1993;6:422-6. 20. Simmons JW, Emery SF, McMillin JN, Landa D, Kimmich SJ. Awake discography: A comparison study with magnetic resonance imaging. Spine 1991;16:216—21. 21. Walsh TR: Weinstein JN, Spratt KF, Lehmann TR, Aprill C, Sayre H. Lumbar discography in normal subjects: A controlled, prospective study. J Bone Joint Surg 1990;72:10Sl-8. 22. Weinstein J, Claverie W, Gibson S. The pain of discography. Spine 1988;13:1344-8. 23. Weinstein JN, Walsh TR, Stratt KF, Lehmann TR, Aprill C, Sayre H. Lumbar discography: A controlled, prospective study of normal volunteers to determine the false-positive rate. Orthop Trans 1989;13:647-8. 24. Whitedoud TS, Seago RA. Cervical discogenic syndrome: Results of operative intervention in patients with positive discography. Spine 19S7;12:313-6. 25. Zucherman J, Derby R, Hsu K, er al. Normal magnetic resonance imaging with abnormal discography. Spine 1988; 13:1355-9. 26. Zucherman J, Derby R: Hsu K, Goldrhwaite N, White A. MRI scan does not replace discography. Orthop Trans 1989; 13:16.
Address reprint requests to Kurt P. Schellhas, MD Center for Diagnostic Imaging 5775 Wayzata Boulevard, Suite 190 St. Louis Park, MN 55416
Point of View James F. Zucherman, MD St. Mary's Spine Center San Francisco, California The authors and their volunteer subjects are applauded for shedding more light on this controversial area. Absolutely speaking, we do not know what a painful discogram means. Clinically, does it mean a solid fusion
will always relieve the pain? Does it mean that there is an abnormal facilitation of nociception, because many disc lesions are not painful on discography? Does it necessarily incriminate the disc wall nociceptors as the
312
Spine • Volume 21 * Number 3 • 1996
pain generator, or can the pain be arising from other structures in the same motion segment that share the same dorsal root ganglion and afferent nerve? When one discogram hurts severely and the adjacent level or levels hurt moderately, is there 'facilitation of otherwise normal adjacent segments by the pathologic one? Nevertheless, clinicians who frequently use the procedure for surgical staging all have the impression of its relative usefulness wrien also considering all.other relevant clinical information. Although the authors do not document statistical significance of their findings, there are definitive trends illustrating the inadequacy of magnetic resonance imaging (MRI) in locating pain generators that do not involve neural compression. Because most patients are plagued with chief complaints of pain, and most do not have neural element compression, understanding the pain generation process has tremendous implications in their care. The implications for the medical—legal—compensation system are equally significant; suffering patients are frequently penalized when the source of their pain is not obvious from simple testing. This paper gives us further evidence that correlation of structural abnormalities (based on MRI and computed tomography [CT] scans} and symptoms may not be present. Discography, as operator dependent and imperfect as it is, fills in the gap for many with severe, refractory disability who would not be treatable otherwise. As the authors suggest, it may also prevent surgical failure
as a result of missing apparent "normal" discs when they are internally damaged and painful in a cervical spine with other gross structural deformities. The value of this paper lies in the authors' success in controlling many of the inherent variables of a clinical study. It shows, as others have, that most cervical discs are morphologically abnormal and some discomfort is usually felt at injection, and that there is a definite painful internal disc response that mimics severe concordant symptoms in symptomatic patients that does not occur in asymptomatic patients. The authors find local anesthetic injection useful in confirming the pathologic level; we have not. The uncontrolled leakage of anesthetic into the epidural space and to other levels, '. and failure to obtain relief in many cases of obviously painful discs that do not leak, has led us to abandon intradiscal local anesthetic as a useful diagnostic tool. In the case of discography-CT scan, we have seen occasional cases of posterior lateral and foraminal protrusion demonstrated only with the use of pressurized contrast, which currently might affect the choice of surgical procedures, and do not agree that the procedure should definitely be dispensed with. In the context of the available information on cervical discography, this paper has contributed much and has done a great service to those patients who really need this procedure to obtain the most successful treatment.
