While the use of discography to diagnose spinal pain syndromes has increased, the practice is not free from controversy. Despite reports of its utility in clinical decision making,3 as well as reports of high sensitivity and specificity,15 including one report of 100% sensitivity and specificity in distinguishing symptomatic from asymptomatic patients with back pain,16 discography is innately subjective and thus can never be completely controlled. This aspect of the therapy relates to the use of pain provocation, which must be concordant with presenting symptoms. As Saal17 notes, most pain-provocative or ablative tests used in the diagnosis of spinal conditions are closely related to the physical examination. In the case of "nonspecific" low back pain created by degenerative lumbar disc disease, the findings from a physical examination are not as clearly defined as those involved in radicular syndromes. While numerous authors have noted the association of certain physical signs with discogenic pain,18-27 only one study in the literature correlates specific physical findings with the results of discography per se: Donelson et al. noted an increased incidence of positive concordant discography in patients who failed to centralize28 according to the criteria of McKen-zie.29 In a specific subset, discography was more likely to be clinically positive (concordant) in patients with annular incompetence. The clinical correlate of this was failure to centralize.
While this is certainly encouraging, it is by no means definitive. One of the key issues in determining the sensitivity and specificity of a particular test is comparison to a recognized and accepted standard of accuracy, a "gold standard." This further complicates the situation for discography, and for invasive spinal diagnostics in general: since a painful joint or disc may have a variable or wide range of anatomical and clinical features that overlap with an asymptomatic structure, a "gold standard" is difficult to define.17 Furthermore, since there is no reliable surgical confirmation of the symptomatic status of a noncompressive degenerative lumbar disc (painful vs painless), surgical confirmation is not a viable option. This inherent level of uncertainty is further compounded by the innately subjective nature of discography as noted earlier.
In an attempt to address false positive findings of lumbar discogra-phy, Carragee et al.6 studied eight subjects, with a total of 24 discograms. None of the patients had a history of low back pain. Patients were scheduled to undergo posterior iliac crest grafting for non-thoracolumbar procedures; 2 to 4 months after the bone graft, patients underwent lumbar discography by a blinded protocol. Fourteen of 24 discs were painful, with two (14.3%) reproducing the pain "exactly." In this case, the pain was referred to the iliac crest bone graft. Based on these results, Carragee et al. concluded that the ability of a patient to separate spinal from nonspinal sources of pain may be questionable. This study is important for several reasons. First, it suggests that discography done under blinded conditions, in accordance with accepted protocol, may in fact not be specific to spinal pathology. Second, it suggests that in patients who were free from other potentially confounding influences (all patients passed a standardized psychometric screening battery prior to the test), significant pain can be produced in a clinically irrelevant setting. Since to be graded positive, pain must be concordant, by definition all those patients in whom no spinal pain source was being evaluated, would have had discordant spinal pain. However, many patients undergoing spine procedures have had iliac crest bone graft harvest. This serves to underscore the possibility that in patients with previous surgery, the findings from discography may be complicated by a confounding variable: a potential pain generator in close anatomical if not physiological proximity.
The technique of discography is of interest as well. While the double-needle technique and multiple blinded injections have become the standard of care, the utility of pressure-controlled discography remains unclear. In a multicenter retrospective study of long-term surgical and nonsurgical outcomes, Derby et al.30 reviewed 96 patients who underwent first discography and then fusion, or continued nonoperative care. These investigators noted no long-term differences in surgical outcome across the entire sample, with the surgical group as a whole doing better than the patients who did not have surgery. In a specific subset, the data suggested that patients with highly pressure-sensitive discs appeared to achieve better long-term outcomes with interbody or circumferential fusion than with intertransverse fusion. For this reason, the authors suggested that there may be a biochemical component to discogenic pain. These results, however, have not been corroborated in prospective studies.
Another requirement for successful discography is the study of a large enough number of discs to permit inclusion of a rostral and, possibly, caudal control. Given that many surgeons have empirically limited arthrodesis to two- or three-level disease in the lumbar spine, the presence of appropriate control levels is critical.
A final note of concern must also be added regarding surgical treatment for discogenic pain. The newer intradiscal therapies are promising but certainly not definitive, and substantial variability exists in surgical outcomes for discogenic pain. Whitecloud and Seago31 reported a 70% rate of clinical success for cervical arthrodesis on the basis of discogra-phy. While Wood et al.32 have noted that thoracic discography may differentiate between symptomatic and asymptomatic degenerated discs (as characterized by the presence of Schmorl's nodes), the optimal surgical treatment of thoracic discogenic pain remains to be identified. In the lumbar spine, a wide variety of success rates have been reported. In one study, an overall success rate of 46% was identified, with a clinical success rate of 96% in the subset that fused solidly.3 Clearly, based on data collected to date, there is no role whatsoever for decompressive surgery in the treatment of discogenic (axial) pain syndromes.
What then are the criteria for "definitive" discography and its use as an indication for reconstructive surgery? Patient selection should be guided by the rigorous criteria.3,20 Strict adherence to technique, including double-needle, multiple blinded injections and identification of rostral and caudal controls, is essential, as is insistence on strict concordance with presenting complaints for a study to be considered "positive." Although the optimal technique of surgical reconstruction has not been definitively identified, the bulk of current literature would probably favor an interbody approach.
Finally, and most important, patient selection is of the greatest importance. Ideally, the patient should be free of confounding organic and psychological pathologies, should have disease limited to one or two levels, and should have reasonable expectations. Perhaps it is in this final area that the thought processes of the diagnostician and surgeon must be most closely aligned.
Was this article helpful?
Deal With Your Pain, Lead A Wonderful Life An Live Like A 'Normal' Person. Before I really start telling you anything about me or finding out anything about you, I want you to know that I sympathize with you. Not only is it one of the most painful experiences to have backpain. Not only is it the number one excuse for employees not coming into work. But perhaps just as significantly, it is something that I suffered from for years.