Discography

Published on 17/03/2015 by admin

Last modified 22/04/2025

Print this page

This article have been viewed 1435 times

CHAPTER 16 Discography

Provocative discography is a diagnostic test sometimes used to evaluate the disc as a potential source of persistent back and neck pain syndromes. In its simplest form, provocative discography is an injection into the nucleus of an intervertebral disc, and the test result is determined by the pain response to this injection. If the injection reproduces the patient’s usual pain, some authors have proposed that the “cause” of the axial pain syndrome can be ascribed to that disc—that is, primary discogenic pain.

In 1948, Lindblom ¹ originally reported discography as a method to identify herniated discs in the lumbar spine by injecting contrast medium into the disc and following the outline of contrast medium into the spinal canal. It was noted as only a secondary consideration of the test that reproduction of the patient’s usual sciatica sometimes occurred during the disc injection. It was observed later that back pain was sometimes reproduced during the injection, as opposed to sciatica. Eventually some clinicians began using the test to evaluate discs as the source of axial pain in patients without radicular symptoms.

Since the early use of discography, it has been unclear whether reproduction of pain with injection indicated that the injected disc is the true primary source of clinical back pain, or whether the injection had simulated the usual pain in an artificial manner. Over time, attempts have been made to determine the specificity of the test and to refine the technique to reduce the risk of false-positive or false-negative results. Still this test remains highly controversial. Even the staunchest proponents of the procedure state that “discography is a test that is easily abused.”² Basic diagnostic test assessment has found fundamental problems with test reliability (i.e., does the test give the same result on repeated testing?) and validity (i.e., does the test prove what it purports to prove?). Also, it has not been shown that using the test improves the outcomes in patients receiving the test compared with patients not receiving the test. More recently, the long-term safety of disc puncture and injection has also been questioned. This chapter discusses the rationale and technique of provocative discography when used in patients with primary axial pain syndromes.

Clinical Context

Back and neck pain are very common, and in most cases determining the “cause” of a specific episode of back or neck pain is unimportant because these symptoms frequently resolve in a short time or do not seriously interfere with function.¹ Provocative discography may be described as representing a tertiary diagnostic evaluation, which should be considered only in a select group of patients.

A primary diagnostic evaluation usually involves screening for serious underlying disease (“red flags”) by history and physical examination aimed at detecting systemic disease, spinal deformity, and neurologic loss. In most patients, these examinations are negative, and nonspecific treatment alone is recommended.

In a patient who does not recover good function in 6 to 12 weeks, a secondary diagnostic survey may be indicated. This follow-up evaluation should identify serious psychosocial barriers to recovery (“yellow flags”) and definitively “rule out” serious conditions that may result in neurologic injury; structural failure; or progression of a visceral disease, systemic infection, or malignant process. Diagnostic tests for serious structural disease, including blood tests and imaging studies, have become so sensitive that these serious conditions are usually identified in the early stages.

Establishing a more specific pathoanatomic diagnosis than “nonspecific back pain syndrome” or “persistent back pain illness” becomes important only if specific therapy directed to common age-related structural changes is considered because of continued serious symptoms and functional loss. At this point, if the primary and secondary evaluations have revealed neither serious structural pathology nor significant confounding psychosocial or neurophysiologic factors, a tertiary diagnostic evaluation may be undertaken. This evaluation may occasionally uncover a clear degenerative cause of symptoms, such as unstable spondylolisthesis or progressive degenerative deformity such as an unstable degenerative scoliosis.

The most common structural degenerative changes (e.g., loss of disc height, loss of nuclear signal, minor facet arthrosis, annular fissures) may be very difficult to reconcile with the severity of apparent symptoms and pain behavior, however, because many people with minimal or no spinal symptoms have similar mild degenerative findings. The question is why do individuals with such benign findings sometimes report severe and persistent pain and impairment? The rationale of provocative discography in the tertiary evaluation is to separate anatomic spinal changes causing serious primary pain illnesses from similarly appearing common degenerative changes that do not cause serious illness. As this chapter shows, it is unclear that this goal is routinely achievable with provocative discography.

Discography Technique

Discography is performed using local anesthetic and mild sedation. The objective is percutaneous injection of a nonirritating radiopaque dye, under fluoroscopic guidance, into one or more intervertebral discs. Ideally, the central portion of the disc, the nucleus, is penetrated by a long fine-gauge needle; this is usually done from a posterolateral approach in the thoracolumbar spine and anterolaterally in the cervical spine. In the lumbar spine, the needle passes posterior to the exiting nerve root and anterolateral to the traversing root. Sometimes a bend of the needle or introducer is required to place the needle accurately, especially at L5-S1.

The passage of the needle in skilled hands should be quick and atraumatic. When the position is verified in two planes using fluoroscopy, the dye is slowly injected into the nucleus of several lumbar discs with the patient blinded to the timing and site of injection. The spread of the dye in the disc is noted on the images, and the patient’s response to injection is documented. The patient is queried at each injection, or at random intervals, whether or not the procedure is painful and is asked to rate the pain against some standardized scale (e.g., 0-5, 0-10, none-to-unbearable). If the injection is painful, the patient is asked to describe the discomfort provoked qualitatively: The injection is usually rated as exactly the same as, or similar to (concordant), or dissimilar to the patient’s usual back or neck pain.

Criteria for Positive Test

In an effort to improve the specificity of discography in diagnosing so-called discogenic pain, some investigators have used additional criteria beyond pain reproduction on injection. The criteria for establishing a positive discogram are controversial. The primary criteria for a “positive” disc injection are pain of “significant” intensity on disc injections (usually defined as ≥6 out of 10 pain scale) and a reported similarity of that pain to the patient’s usual, clinical discomfort (concordant pain). These basic criteria were proposed in the experimental work by Walsh and colleagues in 1990,²⁹ which proposed “significant pain” be defined as 3 out of 5 (or 6 out of 10) on an arbitrary pain thermometer. “Bad pain” was defined as 3 out of 5 pain, and “moderate pain” was described as 2 out of 5 pain. The authors did not stringently define concordance of pain reproduction. Some investigators have proposed additional and sometimes idiosyncratic criteria for positive injections (Table 16–1).

Test Criteria for Positive Result	Positive Test Threshold	Comments
Pain response (intensity)	≥6/10 or 3/5	Subjective and arbitrary scale. No data on reliability. Data on validity in small groups of asymptomatic subjects without psychosocial comorbidity are good (specificity >90%). Data in several studies of subjects with increased psychosocial or chronic pain comorbidity indicate validity in these subgroups is poor (specificity 20%-60%)
	“Bad” pain or worse on pain thermometer
	≥7/10
Qualitative pain assessment (concordant pain)	“Concordant pain” usually including “similar” but not exact pain	Subjective response. Data on reliability are unknown. Data on validity in small study of experimental nondiscogenic low back pain indicate validity is questionable
	“Exact” pain only
Annular disruption	Dye must show fissure to or through outer anulus	Tested only in clinical studies without follow-up to confirm outcome or other “gold standard.” Radiologic reliability best with computed tomography scan after disc injection compared with x-ray alone. Validity of additional criteria as confirming true-positive test unknown; positive injection in discs without annular disruption more common in psychologically disturbed subjects
Control disc injections	“Negative” injection (minimal or discordant pain) required adjacent to proposed “positive” disc	Injections in morphologically normal discs seem to be reliably negative even in subjects with serious psychological distress and no back pain. Reliability in other disc morphology unknown. Validity of this additional criterion as confirming true-positive test unknown
	“Normal” injection (i.e., no pain)
	Some authors insist that adjacent “control disc” must also have grade 3 annular fissure, which is “relatively painless” at equal or higher pressures than “positive disc”
Demonstration of pain behavior	Facial expressions of pain must be observed to confirm verbal pain report	Reliability and validity of this criterion as confirming true-positive test unknown
Pressure-controlled injection	Disc injections should be classified into low (<15 psi or <20 psi) or high (>50 psi) pressures at time of significant pain response; responses at pressures in between are indeterminate	Small outcomes series suggest low pressure sensitive discs are better treated with interbody fusion techniques. Reliability and validity unknown
Volume-controlled injections	“Excessive volume” or speed to injection invalidated injection	Unvalidated concept based on anecdotal evidence. Primary data unavailable to analyze
Maximum one or two positive disc injections	More than one or two positive disc injections invalidates study (all are indeterminate)	Assumption is made that generalized hyperalgesic effect may lead to multiple positive discs around single pain generator
Quantify pain tolerance by response to buffered anesthetic injection	Subjects with poor pain tolerance may not be “ideal” candidates for discography; this feature needs to be detected	It is unclear that pain tolerance to intradermal anesthetic injection is valid test to determine “pain tolerance” in patients with long-standing axial pain
Needles should be inserted from asymptomatic or least symptomatic side	Theoretically this may decrease confusion between injection and insertion pain	Some data suggest this is not an important technique. No “gold standard” confirmation was applied
Any positive disc injection must be repeated with similar outcomes before accepting result as “positive”	Intraprocedure reliability test	No data available on whether this improved or decreased test accuracy

Pain Generator Concept and Provocative Discography

The diagnosis made by a “positive provocative discogram” should indicate that the disc identified is the primary or only cause of the patient’s back pain illness, or the pain generator. This term has proven problematic, however. In a patient with persistent symptoms and a secondary workup with only degenerative findings, the task of identifying a specific isolated pain generator may be formidable. Most patients have multiple findings of disc changes and facet arthrosis, often at different levels. To distinguish which, if any, “degenerative” findings may be definitively established as causing severe back pain illness is a complex problem. Many people have occasional back or neck ache with common activities or episodic axial pain without impairment. The question is not whether any previous or possible future back or neck pain may be coming from a certain spinal structure. Rather, it may be assumed that most people with degenerative change of the axial skeleton may have occasional discomfort from several sites alone or at the same time.

The pertinent question is whether or not a suspected local anatomic structure (e.g., disc, facet, sacroiliac joint) is causing serious, disabling axial pain illness or is only a minor contributor to a generalized pain-sensitivity syndrome (e.g., fibromyalgia), a central pain-processing syndrome, an overuse syndrome related to posture or activity, or other conditions. It is hoped that some diagnostic test can identify whether or not a specific local spinal pathoanatomic structure adequately explains the severity of clinical symptoms. As a matter of practical definition, for a pathoanatomic diagnosis to be clinically relevant requires that the identified pain generator not only be capable of causing some discomfort under any circumstances (e.g., puncture and injection of a disc), but also that this structure is a primary independent cause of the patient’s apparent severe illness.

When only degenerative changes are found, it is controversial whether or not a discrete local pain generator as the cause of serious back pain illness can be commonly identified. Some clinicians believe that serious axial pain and disability can be so multifactorial (mechanical, psychological, social, and neurophysiologic contributors) that it is unreasonable to expect specific diagnostic studies to confirm an anatomic “diagnosis” for axial pain illness in every patient.³^–⁵ Even if a pain generator is suspected, it is unclear how this can be reliably confirmed to be the cause of the patient’s perceived pain, impairment, and disability in the face of complex social, emotional, and neurophysiologic confounders.

Other clinicians believe that identifying a pain generator is central to spine evaluations, is an expectation of patients, and determines the choice of treatments by focusing on the anatomic structure deemed responsible for the pain. In this model, social issues such as disability, litigation, psychological distress, and pain intolerance are believed to be secondary issues to the structural pathology.^1,⁶^–¹² These clinicians generally believe that although the history and physical examination may be helpful in suggesting serious underlying pathology such as infection and tumor, these methods are not helpful in determining the true pain generator among many degenerative structures.

Diagnostic Injections and Modulation of Pain Perception in Axial Pain Syndromes

Provocative discography relies on a patient’s subjective perception and report of pain after a progressive pressurization injection of a disc. Alternatively, a disc may be injected with an anesthetic agent with subsequent documentation of the patient’s subjective pain relief after activities that usually provoke pain. These diagnostic injections seek to identify a primary pain generator by provocative testing (stimulating a potential site of pain as in discography) or by temporary local anesthetic relief. These are subjective tests of pain perception and are subject to the effects of volitional and neurophysiologic modulation at multiple points along the neuraxis.

Many common factors are known to have potential dampening or amplifying effects on the perception of back and neck pain. These factors must be considered when evaluating the validity of diagnoses determined by diagnostic injections.¹³^–¹⁸

Adjacent Tissue Injury

Injury to adjacent tissues may increase the perception of pain in surrounding structures by a local hyperalgesic effect. This is a well-known phenomenon that occurs with any tissue damage; it may amplify pain perception by increasing local inflammatory processes with secondary neurologic sensitization in areas not directly injured, such as the area surrounding a burn or a fracture that is sensitive but without any thermal or mechanical injury. This is an important phenomenon in patients with low back pain with serious disease at one or more levels, which may sensitize the adjacent segments to provocative testing (e.g., a spine that has undergone multiple operations).^13,¹⁹

Local Anesthetic

Local anesthetic injections may decrease the perception of pain at a local site. This is the specific, active effect used in diagnostic blocks. This decrease in pain perception can also be a source of confounding effects if the exact placement of the agent is not well controlled. In addition to this direct local effect, a nonspecific placebo effect and a neurophysiologic modulation effect may occur. A relevant example is the effect of local anesthetic blockade on the perception of painful stimulus along the neuraxis proximal to the injection. A local anesthetic injection in the lower extremity may be perceived as relieving sciatica owing to disc herniation.²⁰ This is not a placebo effect (i.e., the phenomenon is seen only with an active anesthetic agent) but rather an effect on neuromodulation. This effect is important in diagnostic anesthetic blockade as a source of false-positive and false-negative findings.²¹

Tissue Injury and Nociception in Adjacent or Same Sclerotome

Tissue injury having the same or adjacent sclerotomal afferents as lower spinal elements may increase pain sensitivity at any given site. This effect is thought to be due to physiologic and anatomic changes at the level of the dorsal root ganglion or spinal cord ascending tracts. In animal models, single afferent neurons from a dorsal root ganglion may innervate three adjacent discs and a wide range of adjacent structures. This effect is important in considering the specificity of discography at sites of similar embryonic derivation to a known pathologic structure (e.g., nonunion, spondylolisthesis, painful iliac crest bone graft site). The confounding effect of this phenomenon in discography has been experimentally and empirically shown (see later).^13,²²

Chronic Pain Syndromes

Chronic pain syndromes may complicate the evaluation of low back pain syndromes. Chronic pain from regional sites near the lumbar spine (chronic pelvic pain, irritable bowel syndrome, failed hip arthroplasty) or distant to the lumbar spine (chronic neck pain, chronic headache, temporomandibular joint syndrome) may increase pain sensitivity at lower spinal elements. This effect may be regional or global and may be related to neurophysiologic changes at multiple levels along the neuraxis. Preexisting chronic pain syndromes are also associated with depression, narcotic use, and habituation, which have independent pain perception effects. This effect has been shown to have an impact on the intensity of pain intensity from discography in experimental subjects.^13,²²

Narcotic Analgesia

Narcotic medications act at multiple levels to decrease pain sensitivity thresholds, intensity, and affective response. Administration of a narcotic medication may act as a common confounder of diagnostic techniques of any type requiring accurate feedback of pain perception from a patient.^13,^23,²⁴

Narcotic Habituation

Chronic narcotic habituation may act to decrease pain tolerance in the absence of increased narcotic intake. Narcotic habituation decreases endogenous abilities to modulate peripheral nociceptive input. This effect is multifactorial. Chronic narcotic habituation is also associated with depression and sleep disturbances.

Depression, Anxiety, and Somatic Distress

Clinical depression, anxiety disorders, and increased somatic awareness may be seen as predisposing factors to chronic low back pain syndromes or reactions to the pain and disability of chronic low back pain illness or both. In either event, psychological distress usually decreases the pain threshold perception and increases perceived pain intensity and affective response.^5,^23,²⁵ These effects are likely due to central neurochemical changes and systemic effects and have been shown to affect pain responses in discographic evaluation.

Social Imperatives

Overriding social imperatives may result in a decreased pain perception or a dissociation of pain perception and functional loss. A decreased pain perception or even an absence of pain perception despite injury can be seen during some short-term stressful events, such as in accident victims, soldiers in combat, or individuals in certain training environments. Even over long periods, social and cultural factors reinforce low pain reporting and muted or absent pain behaviors.

Social Disincentive

Secondary gain issues may exaggerate pain responses of all types. When the intensity of pain behavior and report is correlated with a real or perceived social benefit or monetary compensation, the reported pain perception and pain behavior may be increased. This situation can have direct effects on provocative testing (e.g., discography) and the need for a specific anatomic diagnosis to establish social validity of an ongoing “sick role.”

Summary

When considering the diagnostic certainty of a possible pain generator in chronic axial pain illness, it is necessary to view the aforementioned confounding factors for contribution to the illness behavior observed (Table 16–2). An injured soldier with facial trauma, after narcotic administration and in the heat of combat, may mask the perception of a significant low back pain injury, which otherwise could be clearly symptomatic. In this case, a bona fide local pain generator results in little pain perception. Conversely, a very minor nociceptive input (common backache) from a disc can be amplified in the case of a patient with multiple chronic pain syndromes, narcotic habituation, depression, and compensation issues (social disincentives). In this case, a common mild backache pain generator is amplified to become a catastrophic illness.

TABLE 16–2 Neurophysiologic Factors Influencing Result of Diagnostic Injections

Modulator of Diagnostic Injection Effect	Type of Effect on Pain Perception at Site of Injection	Diagnostic Effect
Adjacent tissue injury	Increased regional pain perception	Decreased specificity in provocative injection
Local anesthetic	Decreased pain perception at depot site and sometimes in sclerotomal or referral pattern	Decreased specificity in provocative injection
Tissue injury in adjacent or same sclerotome	Increased regional pain perception	Decreased specificity in provocative injection
Chronic pain syndrome	Increased generalized pain perception	Decreased specificity in provocative injection
Narcotic analgesia	Decreased generalized pain perception and affective response	Decreased sensitivity and increased specificity of provocative injections
Narcotic habituation	Increased pain perception and exaggerated affective response	Decreased specificity in provocative injection
Depression, anxiety, and somatic distress	Decreased generalized pain perception and unpredictable affective response	Decreased specificity in provocative injection
Social imperatives	Decreased pain perception, suppressed affective response	Decreased sensitivity and increased specificity of provocative injections
Social disincentives	Specific increased pain reporting and demonstration of pain behavior	Decreased specificity in provocative injection

Evidence for Validity and Usefulness of Provocative Discography

The criteria for an evidence-based evaluation of the validity of diagnostic tests have been described by Sackett and Haynes.²⁶ Four phases of scientific scrutiny and evidence in discography research are shown in Table 16–3. These phases of evidence progress from the simple comparison of testing in subjects known to have a disease with subjects who are completely normal without any signs, symptoms, or morbidity associated with the disease (phase I) to the blinded study of a diagnostic test in determining outcomes in actual clinical therapeutic intervention (phase IV).

TABLE 16–3 Four Phases of Evidence-Based Criteria for Evaluation Diagnostic Tests

Phase of Study	Strategy	Discography Evidence
Phase 1	Diagnostic test compared in subjects with index disease vs. results in complete normals (experimental setting)	Few painful disc injections in completely normal asymptomatic subjects (e.g., normal psychometric testing, normal disc morphology, no chronic pain issues, no compensation issues)
		Phase 1 study examples: Walsh et al, 1990²⁹; Carragee et al, 2000¹⁵
Phase 2	Evaluation of range of test results in subjects with disease (establishes positive result guidelines) compared with known normals	Wide range of pain reactions to injections in asymptomatic subjects depending on psychological status, disc morphology, chronic pain issues, compensation issues. Wide overlap between asymptomatic subjects and patients with presumed discogenic pain.
		Phase 2 study examples: Carragee et al, 2000¹⁵; O’Neill and Kurgansky, 2004²²; Carragee et al, 2006³⁸
Phase 3	Diagnostic test applied in clinical subjects likely to have disease (clinical setting of test application in subjects with similar presentation signs, symptoms, and risk factors)	Poor validity testing subjects with persistent low back pain with known nondiscogenic pain syndromes (e.g., iliac crest pain) or asymptomatic disc pathology (i.e., previous disc surgery). PPV approximately 50% in ideal patient, PPV 50% in typical discography patient (outcome findings)
		Phase 3 study examples: Carragee et al, 1999¹⁴; Carragee et al, 2000³⁵; Derby et al, 2005³⁶; Carragee et al, 2002³⁷; Carragee et al, 2006⁴³
Phase 4	Does having the diagnostic test result improve outcomes compared with management without test result (controlled trial)	Little to no evidence of provocative discography improving outcomes compared with modern diagnostic techniques.⁴⁰ Substantial evidence provocative discography is worse than anesthetic injection alone.⁴² Substantial evidence discography may worsen outcomes in certain at-risk groups⁴⁴^–⁴⁶

PPV, positive predictive value.

After Sackett D, Haynes R: Evidence base of clinical diagnosis: The architecture of diagnostic research. BMJ 324:539-541, 2002.

An example of a phase I diagnostic study is the classic study of discography by Walsh and colleagues ²⁹ in asymptomatic healthy young men without significant degenerative disease or comorbidities associated with chronic low back pain illness (e.g., depression, chronic pain behavior, compensation issues). Discography seemed to perform well in this phase I study, with little pain provocation in the subjects known to have no evidence of disease (1 of 10 subjects [10%, 95% confidence interval 0, 40%] had pain intensity rated “bad”). Phase II and III studies, comparing subjects without low back pain illness but with significant comorbidities, were more problematical, however.^14,¹⁵

Generally, there has been limited high-quality evidence supporting provocative disc injections. Despite the limited evidence, some authors believe that primary discogenic pain is the most common cause of chronic low back pain illness.^11,^12,^27,²⁸ As is shown subsequently, a major constraint in this research has been a failure to use a bona fide “gold standard” for primary discogenic pain causing low back pain illness against which investigators document that the diagnosis suggested by discography is correct.

Validity of Discography

Discography purports to diagnose the presence or absence of a disc lesion responsible for the syndrome of chronic low back pain illness caused by primary discogenic pain. There is no commonly used “gold standard” or criterion to determine who actually has chronic low back pain illness from primary discogenic pain. There are well-accepted standards, however, for who does not—someone with no evidence of significant low back pain. Similarly, someone with new pain resulting from another process (pelvic fracture) does not have “chronic low back pain illness caused by primary discogenic pain.” Provocative discography can be assessed by the results of disc injections in subjects who definitively do not have chronic low back pain illness caused by primary discogenic pain.

Alternatively, a patient’s response to treatment may be considered a surrogate “gold standard” if the treatment definitively removes the pain generator (the disc) and adjustments can be made for surgical and nonspecific limitations of the treatment. The following section describes a series of clinical and experimental studies that have attempted to define the specificity of discography in different at-risk subgroups.

Specificity of Positive Discography: Testing on Subjects with No Axial Pain History

Careful technique and the standardization of discography were believed by many discographers to have reduced the false-positive rate to a negligible level in experienced hands. In 1990, Walsh and colleagues ²⁹ performed a carefully controlled set of discographic lumbar injections in 10 paid volunteers, all asymptomatic young men (mean age 22) with little disc degeneration. Of 30 discs injected in this asymptomatic group, 5 produced “minimum” pain (16.7%), 2 produced “moderate” pain (6.7%), and 1 produced “bad” pain (3.3%). Based on these data, the authors believed the risk of false-positive injections was very low. This study is frequently, and incorrectly, cited to confirm a 0% false-positive rate.

In 1997, a review of one discography practice ²⁵ found cases that seemed to be clinically apparent false-positive cases. These injections were believed to meet full criteria for discogenic pain, with concordant, painful injections and negative control injections. Clinical follow-up revealed other causes of the patients’ back pain illness, however, including spinal tumor, sacroiliac joint disease, and emotional problems. Block and colleagues³⁰ related abnormal Minnesota Multiphasic Personality Inventory (MMPI) testing and Ohnmeiss and colleagues³¹ related abnormal pain drawings with “nonorganic” features, suggesting possible false-positive discographic injections. Other authors performed thoracic³² and cervical³³ injections in subjects asymptomatic for pain in those areas. Significantly painful injections were found to occur in approximately 30% of these volunteers.

Following the Walsh protocol, Carragee and colleagues ¹⁵ examined 30 volunteer subjects with no history of low back pain who were recruited to undergo a physical examination, magnetic resonance imaging (MRI), psychometric testing, and provocative discography. The results showed that little pain was elicited by injection of any anatomically normal disc. Discs with advanced degenerative annular fissuring with dye leakage to the outer (innervated) annular margins were more commonly painful after discography than less degenerative or normal discs. The intensity of the pain reported by the subjects with annular disruption was predicted by the presence of chronic nonlumbar pain and abnormal psychological scores. Only 10% of subjects without any other pain processes had a positive disc injection by the Walsh criteria, but 50% of subjects with nonback chronic pain had at least one positive disc injection.

The interaction between pending compensation claim and discographic pain was also significant in this select group of volunteers. Of the 10 subjects with positive injections, 8 had had contested workers’ compensation or personal injury claims with resulting litigation. Conversely, of 9 subjects with disputed litigation claims, 8 had positive injections (P < .0001).¹⁵ It was not found, however, that all subjects involved in previous work injury claims had similar rates of positive disc injection. A history of an uncontested claim from a past compensation injury and no pending legal action did not predict significant pain on disc injection. Given that no subject in this study stood to have any secondary gain from positive discography, the increased pain reporting in subjects with unrelated but contested compensation claims is intriguing. It is possible that the effect of the prolonged social turmoil associated with a litigation dispute has the effect of diminishing one’s resilience to irritative stimuli. Another explanation could be that persons with abnormally low pain tolerance are more likely to have a legal dispute regarding the significance and damages associated with previous minor injury.

Discographic Injections in Previously Operated Discs

Provocative discography is frequently used to evaluate persistent or recurrent low back pain syndromes in patients who had undergone posterior discectomy. The validity of interpreting painful injections after herniation is unknown despite its common usage. Heggeness and colleagues ³⁴ reported on 83 postdiscectomy patients and found that 72% had a positive concordant pain response on injection of the previously operated disc. This study did not address the possibility of false-positive injections. All positive injections were assumed to be true-positive injections for identifying the source of the patient’s pain.

Using the same methodology developed by Walsh and colleagues,²⁹ a large study of discography in asymptomatic patients after discectomy for sciatica was performed.³⁵ Painful disc injections were frequently seen in the asymptomatic postdiscectomy group. As in previous studies, a higher rate of painful injections was seen in patients with abnormal psychological profiles.

Validity of Concordance Report

Provocative discography is considered positive only when injection elicits the patient’s usual pain in quality and in location. The reliability of the test would be substantially supported if patients could identify the quality of pain coming from a particular disc and differentially compare that sensation with their usual pain.

It is unclear to what extent similar neurologic and behavioral factors may influence the results in provocative discography. It is possible that the disc stimulation in discography may also provoke a “concordant” pain response without actually having located a true pain source. As discussed earlier, there have been reported cases of individuals undergoing discography who were diagnosed as having discogenic pain as the source of their illness on the basis of positive concordant disc injections, but who were subsequently shown to have nonspinal sources for their pain.²⁵

This issue was investigated using an experimental model to determine the response to disc injection in patients known to have nonspinal back pain.¹⁴ Subjects with no history of back pain were recruited who were scheduled to undergo posterior iliac crest bone graft harvesting for nonspinal problems, mainly fracture nonunions or bone tumors. Most of these patients experienced low back and buttock pain from the bone grafting for several months postoperatively, and this pain was in a similar distribution to what is normally considered discogenic lumbar pain. Discography was performed several months after the bone graft harvesting, and subjects were asked to compare the quality and location of disc injection pain with their usual iliac crest pain.

Eight volunteer subjects were studied using the same protocol as the Walsh and colleagues study.²⁹ All subjects had some disc degeneration on MRI, and 24 of the discs were injected. Of the 14 disc injections causing some pain response, 5 were believed to be “different” (nonconcordant) pains (35.7%), 7 were “similar” (50%), and 2 were “exact” pain reproductions (14.3%). The presence of annular disruption was correlated with concordant pain reproduction (P < .05). Of 10 discs with annular tears, injection of 7 elicited “similar” or “exact” pain reproduction to the pain at iliac crest bone graft harvest sites. By the strict criteria for positive discography, four of the eight patients (50%) had positive injections: The pain on a single disc injection was “bad” or “very bad,” and the pain quality was noted to be exact or similar to the usual discomfort. All subjects had a negative control disc. All positive disc injections had annular fissures. Half of the positive disc injections occurred at low pressures (<20 psi).¹⁴

Discography in Subjects with Minimal Low Back Symptoms

The ability of a test such as discography to discriminate a true pain generator disc responsible for causing serious disabling low back pain illness from another disc that causes only trivial or clinically inconsequential backache is critical to the test’s validity in clinical practice. Derby and colleagues ³⁶ performed discography in a group of 16 subjects with occasional or minimal low back pain, none of whom required current medical care or were experiencing disability because of low back pain. Of the 16 subjects, 5 (31%) had a pain response at 5 out of 10 or greater, and 2 (12.5%) had a pain response at 6 out of 10 or greater. The subjects with more frequent benign low back pain had more painful injections. None of these subjects had abnormal psychological profiles, compensation issues, or chronic pain syndromes or had significant secondary gain motivation to underreport pain. This study was confounded by potentially significant methodologic issues making the results open to criticism. These issues include using the investigators’ employees and staff as the subjects of the study.

In another study, the Stanford group performed experimental discography on 25 volunteer subjects with no clinical back pain illness; these volunteers had persistent low backache unassociated with any physical restrictions that was not bad enough to seek medical care.³⁷ All subjects had normal psychological profiles, but half had other chronic pain syndromes that are risk factors for positive injections. In 36% of these subjects with common backache, discographic injection of one or more discs was significantly painful and concordant. All positive discs had annular disruption, and all had negative control discs. By the usual proposed criteria, these were positive disc injections for clinically significant discogenic pain illness. Discs sensitive to low pressure injections were found in 28% of subjects.

Pressure-Sensitive Injections and Discography Validity

In some cases, dye injected at low pressures may cause significant pain. Derby and colleagues ⁸ labeled these “chemically” sensitive discs, as opposed to discs that are painful only on injection with high pressures. These authors theorized that “chemically” sensitive discs are painful because of the exposure of annular nerve endings or nearby neural structures to the leakage of some irritating substances. It is postulated that this pain is incited by chemical leakage from the disc during daily activities. This leakage is thought to be simulated by the disc injections. Low-pressure–positive discs are arbitrarily defined as discs found to be painful at pressures less than 15 psi or 22 psi greater than opening pressures.^8,²² Derby and colleagues⁸ postulated further that disc injections eliciting pain at higher pressures (>50 psi), called “mechanically” sensitive discs, physically distend the anulus and simulate mechanical loading. In these discs, it is presumed that a mechanical deformation of the anulus is the inciting painful event.

The use of pressure measurements has been postulated as a means to decrease the risk of false-positive injections. This assertion would be true if injections were rarely, if ever, positive at low pressures in subjects without true low back pain illness. Previous neurophysiologic considerations suggest that a pain and pressure profile for a given disc lesion may depend on individual pain sensitivity and local pain processes not related to the disc. Hypothetically, the pain and pressure profile may be depicted as shown in Figure 16–1. The presence of the factors enumerated in Table 16–2 thought to have a desensitizing effect would move the curve down and to the right, whereas factors that increase pain sensitivity may move the pain and pressure curve up and to the left.

FIGURE 16–1 Hypothetical responses to pressurization of degenerative disc depending on “pain sensitivity” and “reporting biases” of the patient.

Experimental work has corroborated this hypothetical pain response. Discographic injections have been performed with pressure measurements in asymptomatic or minimally symptomatic volunteers.^15,^29,³⁵^–³⁷ Figure 16–2 shows the proportion of painful injections at low pressures in volunteers with varying risk factors for increased pain sensitization. It seems from these and other data³⁸ that low-pressure injections are more likely positive in subjects with some type of chronic pain state, psychological distress, and, presumably, a generalized sensitization to irritable stimuli. An increased perception of pain at low-pressure injections seems to affect the pain response even when the chronic pain state is not in the low back region.

FIGURE 16–2 Discography testing in asymptomatic subjects with varying risk factors. Proportion of painful disc injections and painful injections at low pressures seems to increase with increasing risk factors.

(Data from references 15, 20, 29, and 36.)

Evidence That Discography in Clinical Practice May Improve Outcomes

Many case series report that provocative discography is helpful in management of patients with chronic low back pain illness. These are uncontrolled studies, however, and the relationship of discography findings to clinical outcome after surgery is speculative: Good outcomes when encountered may be the result of nonspecific effects, natural history of the condition independent of diagnosis or treatment, scrupulous patient selection, or confounding findings on standard imaging studies. In a retrospective literature review, Cohen and Hurley ³⁹ compared outcomes of spinal fusions in studies that included discography with fusion outcomes without preoperative discography. The outcomes were not significantly different.

In the era before routine MRI use, Colhoun and colleagues ⁴⁰ retrospectively compared a series of fusions planned with and without preoperative discography. Evaluation methods of these patients having surgery in the 1970s and early 1980s did not include dynamic radiographs, MRI, or computed tomography (CT). The authors reported better results in the discography group. The two groups were not similar at baseline, however, and potential biases resulting in some patients having preoperative discography and others not having it were not examined. In the era before contemporary imaging techniques, this study suggested that discography may assist in the evaluation of patients before surgery. This is an extremely rare clinical scenario today and has little applicability to current evaluation protocols.

Madan and colleagues ⁴¹ did a retrospective review of consecutive patients undergoing spinal fusion performed by the same surgeons, with and without preoperative discography. The two groups seemed well matched for demographic, psychometric, and radiographic features. At a minimum of 2-year follow-up, there was no significant difference in outcome between the two groups. The addition of discography to radiographs and MRI did not improve outcomes compared with radiographs and MRI alone.

A more recent randomized clinical trial compared outcomes of subjects having single-level fusion based on preoperative evaluation using provocative discography with subjects having an anesthetic disc injection.⁴² These were in some respects subjects with best-case scenarios (no psychological distress, no depression, no workers’ compensation cases, and no traffic accident litigants). The discography was performed using low-pressure injections. The outcomes in the discography group were uniformly worse than the group using an anesthetic block to determine fusion (Fig. 16–3). A phase IV evaluation (based on evidence-based criteria as described by Sackett and Haynes²⁶) of discography has not been performed to date (see Table 16–3).

FIGURE 16–3 Randomized clinical trial by Ohtori and colleagues comparing the proportion of good outcomes for pain and function after single-level spinal fusion in patients selected by best clinical judgment and provocative discography with best clinical judgment plus anesthetic disc injection. These were “best-case” scenario subjects in many respects: no workers’ compensation cases, no road traffic accident litigation cases, no high somatic distress cases, no depression cases, all selected by very experienced spinal surgeons in Japan. Outcomes are clearly inferior when discography is used to select patients for fusion.

(Data from Ohtori S, Kinoshita T, Yamashita M, et al: Results of surgery for discogenic low back pain: A randomized study using discography versus discoblock for diagnosis. Spine 34:1345-1348, 2009.)

Clinical Outcome as a “Gold Standard” in Provocative Discography

From the evidence reviewed in this chapter, discography has been shown to be frequently positive in asymptomatic subjects and in subjects with pelvic pain owing to iliac crest harvesting and has been shown to be frequently fully concordant in subjects with clinically insignificant backache. These findings suggest there is limited experimental evidence to support the premise that discography can accurately identify clinically significant lesions responsible for a patient’s chronic low back pain illness. Direct assessment of a positive test against an accepted “gold standard,” confirming a true-positive result, has not been performed.

A common empirical “gold standard” would involve comparing the test results with clinical surgical outcomes, assuming an excellent clinical outcome would confirm a true-positive test. There is concern, however, that an excellent clinical result may overestimate the number of true-positive results because of a placebo or nonspecific effect of spinal fusion or other intervention. There is also concern that clinical outcomes may underestimate the number of true-positive tests because the ability to achieve outstanding results is limited by patient-specific variables (psychological distress or social issues preventing recovery despite surgical cure of the lesion) or by operative morbidity or technical limitations, which, even in the best of cases, cannot achieve 100% success even with an accurate diagnosis.

An attempt was made to control these variables (patient-specific variables and operative comorbidities) in a prospective controlled study of spinal fusion for presumed diagnoses of unstable spondylolisthesis versus discogenic pain diagnosed by discography. Identical operative techniques were used, and patients had no psychosocial comorbidities.⁴³ Both groups included only highly selected patients with 6 to 18 months of severe low back pain, normal psychological testing, no previous or concomitant pain syndromes, and no workers’ compensation or personal injury claims. All patients had either positive discography at one level only and at low pressures (<20 psi) or unstable spondylolisthesis by strict radiographic criteria. All patients had been working full-time before their back problem, and no patient was taking daily narcotic medications.

Both groups underwent an anterior spinal fusion with posterior instrumentation and fusion. Two years after surgery, only 27% of patients in the discography group met stringent criteria for clinical success compared with 71% of the spondylolisthesis group. Success was defined as full return to work and recreational activities, pain scores on a visual analog pain scale (VAS) less than 2, Oswestry Disability Index score less than 15, and no daily medications for back pain. Even using less rigorous outcome measures, 43% of the discography group compared with 91% of the spondylolisthesis group reported at least moderate improvement. Even after controlling for operative morbidity, the maximum proportion of true-positive discograms in a best-case scenario (i.e., assuming normal psychometric testing, no other chronic pain history, no compensation issues or litigation, and single-level degeneration) was 40% to 60%, with a false-positive rate of approximately 50%.

For less “ideal” patients, provocative discography may be an extremely poor tool to select appropriate operative candidates. Freedman and colleagues,⁴⁴ using CT and provocative discography to select a wide range of typical low back pain subjects for an intradiscal electrothermal therapy trial (including patients with psychometric distress and compensation claims), found no improvement at all compared with control subjects. Pauza and colleagues,⁴⁵ in a similar intradiscal electrothermal therapy trial but excluding subjects with psychological abnormality, workers’ compensation claims, or litigation claims, found only slightly better outcomes: Most subjects did not have improvement better than expected by placebo.

Even more striking, Derby and colleagues ⁴⁶ found such poor outcomes of spinal fusion after provocative discography in patients with abnormal mental component scores (MCS) on the SF-36 that the discography seemed to preselect patients who were extremely unlikely to have a satisfactory outcome. These results, illustrated in Figure 16–4, may be substantially worse than using alternative patient selection strategies without discography (e.g., radiographs, MRI, patient interview, or psychological screening).

FIGURE 16–4 Outcomes of spinal fusion when discography was used in patient selection. Subjects with positive discogram in setting of abnormal mental component scores in SF-36 were highly unlikely to improve or reach even minimum clinically important change in physical outcomes. In contrast, subjects with more normalized mental component scores had significantly better improvement in outcomes (P < .005). In this study, positive provocative discography result in clinical subset of psychologically distressed patients seems to select patients unlikely to improve with surgical treatment.

Complications

Although there are many potential complications of any invasive procedure, several potential complications of discography warrant specific discussion, as follows:

1 Infection: There is a small but definite risk of discitis after percutaneous puncture and injection. The absolute risk is difficult to calculate, but modern methods likely limit this risk to much less than 1%. Double needle techniques for insertion, less irritating dye, and intravenous or injectate antibiotics all have been postulated to decrease the infection risk.

2 Prolonged pain episode: Occasionally, patients may experience a prolonged episode of pain after a disc injection. One reason given for this phenomenon is the hypothetical displacement of fibrous repair over annular fissures owing to disc pressurization.² Other work has shown that 40% of subjects with psychological distress at the time of injection can have markedly increased back pain for 1 year after discography. This effect was not seen in subjects with normal psychological profiles.⁴⁷

3 Misleading diagnosis resulting in inappropriate or nonproductive invasive treatments: As discussed previously, subjects with one or more risk factors for false-positive testing may be misdiagnosed as having primary discogenic pain as the cause of their persistent low back pain illness. Patients with abnormal psychometric testing undergoing surgery based on this test are extremely unlikely to have substantial benefit from disc-directed interventions (see Fig. 16–4) and are exposed to the hazards and morbidity of these procedures.⁴⁶

4 Accelerated disc degeneration: In animal models, disc puncture with a needle has provided a reliable model to initiate rapid disc injury with structural changes similar in some respects to naturally occurring disc degeneration. Working with a large animal model, Korecki and colleagues ⁴⁸ showed that relatively minor disruption in the disc from even a 25-gauge needle puncture injury had “immediate and progressive mechanical and biologic consequences with important implications for the use of discography….” Similarly, Nassr and colleagues⁴⁹ showed that needle puncture in cervical discs during cervical spinal surgery localization radiographs was apparently associated with a threefold risk of rapid disc degeneration.

Carragee and colleagues ⁵⁰ performed a prospective, match-cohort study of disc degeneration progression over 10 years with and without baseline discography. The investigators performed a protocol MRI and L3-4, L4-5, and L5-S1 discography examination at baseline in 75 subjects without serious low back pain illness. The investigators enrolled a matched group at the same time and performed the same protocol MRI examination. Subjects were followed for 10 years. At 7 to 10 years after baseline assessment, eligible discography and control subjects underwent another protocol MRI examination. MRI examinations were scored for qualitative findings (Pfirrmann grade, herniations, endplate changes, and high-intensity zone). Loss of disc height and loss of disc signal were measured by quantitative methods (Fig. 16–5). The investigators found that modern discography techniques with small-gauge needle and limited pressurization resulted in accelerated disc degeneration, disc herniation, loss of disc height and signal, and development of reactive endplate changes compared with matched controls.

FIGURE 16–5 A-D, Progression of disc degeneration in matched cohorts of subjects, discography (Disco) versus nondiscography controls. Baseline versus 10-year follow-up MRI studies were compared for progression of Pfirrmann grade (A), development of new disc herniations (B), loss of disc height and nuclear signal (C), and development of new Modic findings or high-intensity zones (D). In all parameters, degeneration was greater in discography group.

(Data from Carragee EJ, Don AS, Hurwitz EL, et al: 2009 ISSLS Prize Winner. Does discography cause accelerated progression of degeneration changes in the lumbar disc: A ten-year matched cohort study. Spine 34:2338-2345, 2009.)