Chronic low back pain (LBP) is a common health issue in the United States with significant economic consequences for the afflicted individuals. Available evidence suggest that the intervertebral disc is one of the most common sources of LBP, accounting for approximately 40 % of cases. The pathological basis for discogenic LBP might entail full-thickness radial tears of the annulus fibrosus. While magnetic resonance imaging (MRI) can image the internal morphology of the intervertebral disc, Zhou and Abdi (2006) reported that relative low sensitivity (26.7 % to 59 %) as well as high false-positive (24 %) and false-negative (38 %) rates decrease the value of MRI in screening for the existence of internal disc disruption as a cause for discogenic LBP. Furthermore, many radiographical abnormalities observed on MRI in patients with LBP are also seen in asymptomatic individuals.
While MRI can detect disc degeneration, it can not confirm if a disc is symptomatic and responsible for the patient's pain syndrome. Lumbar discography (also known as lumbar provocative discography and provocative lumbar discography) is a test that is used to ascertain if a disc is painful on injection. It is an invasive procedure that entails the injection of radiopaque contrasting materials (1 to 3 ml) into the intervertebral disc followed by computed tomography (CT) to examine disc abnormality. Discography can provide radiographical evaluation of the integrity of the nucleus pulposus and annular rings to determine tears or other lesions that could be the cause of LBP. It can measure disc nociception -- a normal disc should not cause pain when injected; however, a disc that is physiologically compromised can mimic the pain experienced by the patient.
Discography is usually carried out when other diagnostic tests have failed to identify the cause of LBP. However, its use is still controversial. Carragee and colleagues (2006 investigated the hypothesis that provocative discography by strict criteria accurately identifies a LBP illness due to a primary disc lesion. Over a 5-year period using a strict enrollment protocol, 32 patients with low LBP and a positive single-level low-pressure provocative discogram, underwent spinal fusion. Subjects with known patient selection co-morbidities were excluded. Generic surgical limitations/morbidity were controlled by comparison to the clinical outcomes of a strictly-matched cohort of 34 patients having a well-accepted single-level lumbar pathology (unstable spondylolisthesis). Treatment success was compared between groups. In the control-spondylolisthesis group, 23 of 32 patients (72 %) met the highly effective success criteria compared with 8 of 30 in the presumed discogenic pain cohort (27 %). The proportion of patients who met the "minimal acceptable outcome" was 29 of 32 (91 %) in the spondylolisthesis group and 13 of 30 (43 %) in the presumed discogenic pain group. Adjusting for surgical morbidity and dropout failure, by either criteria of success, the best-case positive predictive value of discography was calculated to be 50 to 60 %. The authors concluded that positive discography was not highly predictive in identifying bona fide isolated intradiscal lesions primarily causing chronic serious LBP illness in this first study comparing discography results to a gold standard.
Wichman (2007) stated that lumbar discography has been used to identify the source of LBP when non-invasive imaging (e.g., MRI) does not reveal morphologic abnormality consistent with symptoms. Controversy regarding the usefulness of discography has been ongoing for over 50 years. Modern advancements with imaging and technique still have not been sufficient to justify the practicality of this procedure for standard use. Based on review of current literature, pain provoked by discography of normal appearing discs on MRI were likely due to internal disc disruption, increased pain sensitivity, and false positive result with chronic pain, psychological state, central hyperalgia, and technical difficulty of the procedure. These causes of positive pain provocation are not amendable to invasive treatment. In these cases, an invasive diagnostic procedure to identify problems best treated with conservative management is not practical. The author concluded that there is no clear evidence-based purpose for discography in the diagnosis and treatment of LBP.
On the other hand, there are studies that found discography to be useful in the management of patients with LBP. Yuan and associates (2006) examined the clinical importance of discography and CT after discography (CTD) in lumbar disc diseases, and compared CTD with MRI. A total of 265 patients (177 males and 79 females, 298 discs) with back and leg pain underwent discography followed by CT 2 hours later. All discs were classified into 7 types according to the findings by discography and CTD. Comprehensive MRI pictures were available in 237 of the 265 patients (with 257 discs); they accepted the MRI classification and were compared with the CT-discographic findings. A total of 245 patients underwent operation, among which 247 discs showed visual lesions during the procedure and 238 cases (96.4 %) had their disc diseases diagnosed accurately through discography and CTD. Considering the positive screening rate for the discogenic abnormality, the difference was of statistical significance between the CTD and MRI groups (p < 0.01). MRI was considered to have limited value in disc diseases with passive discographic finding. These investigators concluded that discography is a safe and effective assistant diagnostic tool, which can gather special information regarding disc abnormality. The matched-group study showed that MRI and CTD can not replace each other. MRI may act as the screening tool for disc diseases, but is inferior to CTD, especially for the contiguous disc structure in the spinal fixation procedure.
Hao and associates (2006) examined the diagnostic value of lumbar discography in discogenic LBP and the effects of intervertebral fusion surgery. A total of 45 patients with 101 discs underwent discography, 360 degree fusion manipulation were performed on 18 discography abnormal cases, 27 cases were treated conservatively. Discography, MRI and provocative pain were observed and all the cases were followed-up. Twenty-one cases showed positive provocative pain (21/45, 47 %), and 21 discs of 101 were concordant discography (21/101, 21 %). All cases were followed for an average 16 months (15 to 23 months), the satisfactory rate was 83 % (15/18) in the surgery group and 41 % (11/27) in the conservative group. The authors concluded that the discography is moderately sensitive in the diagnosis of discogenic LBP. Furthermore, the short-term follow-up reveals that operative group has better pain relief than the conservatively-treated group.
Kosharskyy and Rozen (2007) stated that diagnosing discogenic LBP is essential to successful treatment. Traditionally, this should be a LBP in a "band-like" distribution without radiculopathy that is worse in the morning, worse with Valsalva, and aggravated by standing in flexion. These researchers noted that provocative discography with manometric monitoring plays an important role in aiding the diagnosis.
Buenaventura and associates (2007) evaluated the diagnostic accuracy of discography with respect to chronic spinal pain. Study inclusion/exclusion criteria were based on the modern practice of discography. Selected studies were then subjected to 2 rating instruments for diagnostic accuracy studies (AHRQ and QUADAS). Specific data were then culled from these studies and tabulated. Evidence was then classified into 5 levels: (i) conclusive, (ii) strong, (iii) moderate, (iv) limited, or (v) indeterminate. Evidence is strong for the diagnostic accuracy of discography as an imaging tool. Evidence is also strong for the ability of discography to evoke pain. There is strong evidence supporting the role of discography in identifying that subset of patients with lumbar discogenic pain. There is moderate evidence supporting the role of discography in identifying a subset of patients with cervical discogenic pain. There is limited evidence supporting the role of discography in identifying a subset of patients with thoracic discogenic pain. The authors concluded that discography is a useful imaging and pain evaluation tool in identifying a subset of patients with chronic spinal pain secondary to intervertebral disc disorders.
The North American Spine Society's position statement on lumbar discography (Guyer 2003) stated that most of the literature supported the use of discography in select situations. Indications for discography include (i) evaluation of patients with abnormal discs to help assess the extent of abnormality or correlation of the abnormality with the clinical symptoms, which may include recurrent pain from a previously operated disc and lateral disc herniation, (ii) evaluation of patients with persistent, severe symptoms in whom other diagnostic tests have failed to reveal clear confirmation of a suspected disc as the source of pain, (iii) evaluation of patients who have failed surgical intervention to ascertain if there is painful pseudarthrosis or a symptomatic disc in a posteriorly fused segment and to help assess possible recurrent disc herniation, (iv) evaluation of patients before spinal fusion to determine if the lumbar discs within the proposed fusion segment are symptomatic and to ascertain if discs adjacent to this segment are normal, and (v) evaluation of candidates for minimally invasive surgical intervention to confirm a contained disc herniation or to investigate dye distribution pattern before chemonucleolysis or percutaneous procedures.
The American Association of Neurological Surgeons/Congress of Neurological Surgeons' guidelines for the performance of fusion procedures for degenerative disease of the lumbar spine (Resnick et al, 2005) stated that discography is not recommended as a stand-alone test for treatment decisions in patients with LBP. It should not be attempted in patients with normal lumbar MRI. Discography appears to have a role in the evaluation of patients with LBP, but it is best limited to the evaluation of abnormal interspaces identified on MRI, the investigation of adjacent-level disc disease, and as a means to rule out cases of non-organic pain from surgical consideration.
While discography may play a role in the evaluation of patients with discogenic LBP, its role in cervical and thoracic discogenic pain is not firmly established. Cervical discography has been employed to aid in ascertaining the specific level or levels causing cervical pain and, potentially, which levels to fuse (Zheng et al, 2004; Wieser and Wang, 2007). The American College of Radiology's panel on musculoskeletal imaging for chronic neck pain (Daffner et al, 2005) did not recommend discography for the evaluation of patients with chronic neck pain. Furthermore, the American Society of Interventional Pain Physicians' evidence-based practice guidelines in the management of chronic spinal pain (Boswell et al, 2007) stated that among the diagnostic interventions, the evidence is strong for lumbar discography, whereas, the evidence is limited for cervical and thoracic discography.
Although lumbar discography may play a role in evaluating patients with discogenic LBP and aids in selecting patients for surgical intervention, it does not appear to improve surgical outcomes. Madan et al (2003) reported on their findings of 73 consecutive patients who underwent postero-lateral interbody and posterior spinal arthrodesis for discogenic LBP that is refractory to non-operative management. Chronologically, the first 41 patients (group A) were indicated without discography, whereas the remaining 32 patients (group B) had been indicated only if their pain had been reproduced during disc injection. The 2 groups were similar in demographical, psychometrical, and radiological parameters. Average follow-up times for group A and group B were 2.8 years and 2.4 years, respectively, both with a 2-year minimum. Using modified Oswestry scoring, group A and group B patients had satisfactory outcomes of 75.6 % and 81.2 %, respectively. This difference was neither statistically significant nor suggestive. The authors concluded that pre-operative lumbar discography did not improve surgical outcomes after circumferential fusion for discogenic LBP.
Cohen and associates (2005) provided a clinical overview of lumbar discography and discogenic LBP, with special emphasis on determining the accuracy of discography and if the procedure improves outcomes for surgery. Based on a large number of comparative studies, plain discography is less accurate than MRI in diagnosing lumbar herniated nucleus pulposus and comparable or slightly more sensitive in detecting degenerative disc disease. For disc degeneration, discography followed by CT remains the gold standard for diagnosis. There are very few studies comparing surgical outcomes between patients who have undergone pre-operative provocative discography and those who have not. What little evidence exists is conflicting. Before disc replacement surgery, approximately half the studies have used pre-operative discography. A comparison of outcomes did not reveal any significant difference between the 2 groups, but none of the studies was controlled, and they used different outcome measures, follow-up periods, and surgical techniques. The authors concluded that although discography, especially when combined with CT, may be more accurate than other radiological studies in detecting degenerative disc disease, its ability to improve surgical outcomes has yet to be proven.
Willems et al (2007) assessed if the pre-operative status of the adjacent discs, as determined by provocative discography, has an impact on the clinical outcome of lumbar fusion in patients with chronic LBP. In 197 patients with an equivocal indication for lumbar fusion (2/3 were patients with prior spine surgery), the decision for either lumbar fusion or conservative management was determined by a temporary external transpedicular fixation trial. During the diagnostic work-up, all patients had undergone provocative discography that included the assessment of the discs adjacent to the intended fusion levels. The individual changes in pain on a visual analog scale, assessed before treatment and at follow-up, and patient satisfaction were the outcome measures. In the 82 patients who underwent a lumbar fusion, no difference in outcome was found between those patients with degenerative or symptomatic discs adjacent to the fusion and those with normal adjacent discs. The authors concluded that in this cohort study of chronic LBP patients with an uncertain indication for lumbar fusion, the pre-operative status of adjacent levels as assessed by provocative discography did not appear to be related to the clinical outcome after fusion.
While discography may demonstrate the connection between an abnormal image and the source of LBP, it does have notable related risks. The most serious complication of lumbar discography is discitis, a post-discography infection with an incidence of 1 to 4 % (Osti et al, 1990). It is thought to be the consequence of bacterial penetration of the disc by a needle contaminated with skin flora. The use of prophylactic antibiotics has been advocated, although there is question regarding the effectiveness of this approach and possible adverse effects on disc cells. Willems and colleagues (2004) studied the incidence of post-discography discitis without the use of prophylactic antibiotics in a consecutive patient group. Additionally, a systematic literature review was performed using strict criteria: (i) discography was performed by means of a double-needle technique, (ii) complications such as discitis were specifically looked for at follow-up, and (iii) the exact numbers of patients and those lost to follow-up were reported. The clinical results of 200 patients with 100 % follow-up for a minimum period of 3 months showed no case of discitis. In the literature review, 10 studies were selected. Nine studies without prophylactic antibiotics reported an overall incidence of 12 cases in 4,891 patients (0.25 %) or 12,770 discs (0.094 %). The only study with prophylactic antibiotics (127 patients) showed no case of discitis. The authors concluded that regarding the small number of patients in the only study in which antibiotics were used and the overall low incidence of post-discography discitis, not enough evidence was found that prophylactic antibiotics can prevent discitis. It was concluded that in lumbar discography by means of a double-needle technique without prophylactic antibiotics, the risk of post-discography discitis is minimal and there is not insufficient evidence to justify the routine use of prophylactic antibiotics.
An assessment by the Centers for Medicare & Medicaid Services (CMS, 2008) has concluded that the effect of discography on clinical outcomes is uncertain. The CMS explained that uncertainty exists in the selection of appropriate patients for the various procedures to treat low back pain that is suspected to be caused by the disc. CMS noted that patient symptoms alone do not accurately point to the disc as the source of pain. The CMS assessment explained that identifying the pain generator in LBP is challenging due to the anatomic complexity of the spine and poor understanding of neurophysiologic mechanisms of pain sensation.
The CMS assessment stated that the literature maintains that for clinicians to identify patients appropriate for back pain procedures, positive provocative discography and either MRI and/or CT indicating IDD (or disease) should be relied on. Citing an assessment by the Danish Centre for Evaluation of Health and Technology (DACETA, 2003), the CMS assessment stated "[h]owever, the value of discography as a diagnostic tool in the identification of discogenic pain is controversial. The benefit of discography and abnormalities in imaging studies in guiding a patient’s treatment for back pain is questioned." The CMS assessment cited evidence that patients with nonspinal back pain and asymptomatic patients can have positive discography. The CMS assessment also cited evidence that, in patients with LBP, provocation discography can be positive in the absence of imaging structural or physiologic abnormalities. The CMS assessment also cited evidence that showed that, while imaging studies can demonstrate structural and physiologic abnormalities, there is no correlation with either the presence or severity of LBP. The assessment noted that, in fact, patients can have annular tears on MRI or CT and be asymptomatic. The CMS assessment quoted Chou et al (2005), who stated, “Although a bright, focal increase of T2-weighted magnetic resonance signal in the posterior annulus (the so-called high intensity zone lesion) is correlated with annular tears, it is not correlated with low back pain.” The CMS assessment also found that Slipman et al (2001) showed no statistical correlation between the side of the patients’ concordantly painful annular tear and the side of the patients’ LBP during discography. The CMS assessment found that these findings directly question if the annular tear, whose natural history is unknown, is the pain generator being tested during discography. The CMS assessment concluded that, "[i]n summary, the effect that discography has on patient outcomes (mediated through the choice of therapy) is uncertain and no specific anatomic lesion has been proven to be the source of discogenic low back pain."
An assessment of provocative discography prepared for the Washington State Health Technology Assessment Program by the ECRI Institute (Schoelles et al, 2007) reached similar conclusions about the lack of reliable evidence of the clinical utility of this test. The assessment concluded that, "[d]ue to the lack of evidence of sufficient quality, we drew no conclusions about whether performing discography influences clinical outcomes." The assessment identified 3 studies that they judged to be of "low quality" that addressed prediction of surgical success and outcomes; 2 of these studies were found to be not favorable. They identified no study that addressed impact on therapeutic decisions.
The Washington State Health Technology Clinical Committee (HTCC, 2008) found that current evidence does not demonstrate that provocative discography produces reliable results, even though expert opinion evidence supports the use of discography to rule out surgery. The Committee found that the available evidence on specificity is of low quality and focuses on a secondary result of whether the same reader can later read the test in the same way, instead of the more substantive data on whether administering the test at different times produces the same results. The second issue is that the primary outcome relied upon in the test is the replication of normal pain. The Committee observed that there is wide clinical debate on the ability to measure accurately, and the meaning of, the subjective pain response. The relevance is made more unclear by findings that there is no established clinical case definition for degenerative disc disease other than radiographic and other imaging descriptions. The issue of false positives and reliability (percent agreement) are also a key concern raised by several studies, including studies that demonstrated a high rate of positive discography findings in asymptomatic individuals. The Committee noted that patients may be subject to additional tests and risks of invasive therapies unnecessarily. The Committee found that no professional guideline recommends provocative discography as a stand alone or pre-operative diagnostic test. MRI is recommended as the diagnostic test of choice. The Committee noted that several guidelines do not recommend the use of provocative discography at all, while several advocate its use in addition to other tests.
Carragee et al (2009) compared progression of common degenerative findings between lumbar discs injected 10 years earlier with those same disc levels in matched subjects not exposed to discography. A total of 75 subjects without serious LBP underwent a protocol MRI and an L3/4, L4/5, and L5/S1 discography examination in 1997. A matched group was enrolled at the same time and underwent the same protocol MRI examination. Subjects were followed for 10 years. At 7 to 10 years after baseline assessment, eligible discography and controlled subjects underwent another protocol MRI examination. Graders of MRI, blind to group designation, scored both groups 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. Well-matched cohorts, including 50 discography subjects and 52 control subjects, were contacted and met eligibility criteria for follow-up evaluation. In all graded or measured parameters, discs that had been exposed to puncture and injection had greater progression of degenerative findings compared to control (non-injected) discs: progression of disc degeneration, 54 discs (35 %) in the discography group compared to 21 (14 %) in the control group (p = 0.03); 55 new disc herniations in the discography group compared to 22 in the control group (p = 0.0003). New disc herniations were disproportionately found on the side of the anular puncture (p = 0.0006). The quantitative measures of disc height and disc signal also showed significantly greater loss of disc height (p = 0.05) and signal intensity (p = 0.001) in the discography disc compared to the control disc. The authors concluded that modern discography techniques using small gauge needle and limited pressurization resulted in accelerated disc degeneration, disc herniation, loss of disc height and signal and the development of reactive endplate changes compared to match-controls. Careful consideration of risk and benefit should be used in recommending procedures involving disc injection.
A practice guideline from the American Pain Society on LBP (Chou et al, 2009) stated that provocative discography is not recommended for diagnosis of patients with chronic, non-radicular LBP.
Functional anesthetic discography (FAD) is a new diagnostic procedure that involves the injection of a local anesthetic (e.g., lidocaine or bupivacaine) directly into 1 or more suspected discs using a balloon-anchored catheter for confirming the presence of injured discs as the source of a patient's LBP. In particular, FAD entails a functional examination; it relies on the patient's induction of pain during active patient movement, which is far different than the traditional discography. When the patient performs the movement or position that causes pain, a local anesthetic is then injected into the disc in hopes of alleviating the pain. If the injection relieves the patient's LBP, the disc can be further evaluated for potential treatment. If the injection fails to relieve the patient's LBP, the physician can investigate other possible causes of pain. However, there is insufficient scientific evidence regarding the clinical value of FAD.
In a case-series study, Alamin et al (2008) presented the findings of 3 patients in whom FAD was used for the evaluation of presumptive discogenic LBP. In 1 patient, FAD was confirmatory of the results of provocative discography; while the test results were divergent in the remaining 2 patients. The authors concluded that further study of FAD will allow more definitive recommendations with regards to the validity and utility of this new technique.
In a meta-analysis, Kapoor et al (2010) estimated the incidence of discitis after cervical discography, delineated the consequences of discitis, and identified factors that may influence this complication. Studies pertaining to cervical discography were identified by a literature review and bibliographic search. These were screened for inclusion into the meta-analysis by 2 reviewers. Data were collected on a wide range of clinical and demographic variables including age, gender, morbidities, number of patients, number of discograms, use of prophylactic antibiotics, type of surgical prep, number of needles used, and the number of patients and discs infected. Primary data were used to calculate the incidence of discitis per patient and per disc. A total of 14 studies were included in the analysis. Both procedural details and demographical information on patients were missing from 8 studies. The mean age of patients ranged from 41 to 47 years, and gender distribution varied greatly. Antibiotics use was reported in 3 studies. Cervical discography was complicated by post-procedural discitis in 22 of 14,133 disc injections (0.15 %) and 21 of 4,804 patients (0.44 %). Only 1 patient suffered from an infection at more than 1 spinal level. The authors concluded that the rate of discitis after cervical discography is relatively low. This can perhaps be further decreased by the use of prophylactic intra-discal antibiotics. They stated that should the ability of cervical discography to improve surgical outcomes be proven, the fear of discitis should not preclude performance of disc provocation.
In a prospective case-series study, Alamin et al (2011) compared the results of standard pressure-controlled provocative discography (PD) to those of the FAD in a series of patients presenting with chronic LBP and considering surgical treatment. A total of 52 patients presenting with chronic LBP (mean age of 45 years with a range of 24 to 70; 24 men and 28 women; and 25 % in workman's compensation program) were included in this study. Outcome measures included Oswestry, visual analog scale for back pain, distress and risk assessment method psychometric analysis, demographic data, MRI scan of lumbar spine. Other outcome measures entailed: during PD -- pressurization at pain, pain level, and concordancy; during FAD -- position or activity used to elicit typical pain, baseline pain level before injection, during injection, at 5, 10, and 20 mins after the injection, and substance injected. Standard pressure-controlled PD was performed, followed by (in positive cases or in patients with clinical features and imaging studies felt to be highly suggestive of symptomatic disc degeneration) the FAD test-an assessment of the response to injection of a low-dose of local anesthetic into the disc during a position productive of the patient's typical pain. Discordant results of the 2 tests were noted in 46 % of the patients in the series. Of them, 26 % of patients with positive PD had negative findings on the FAD test; 16 % had positive findings at a single-level only, whereas PD had been positive at 2 or more levels; 4 % had new positive findings on the FAD test. The authors presented the results of a new diagnostic technique in 52 patients with chronic LBP presumed discogenic in origin that was designed to help differentiate between symptomatic and asymptomatic disc degeneration. The findings of the test differed from those of standard pressure-controlled PD in 46 % of the cases reported on here. They stated that further studies are needed to demonstrate the clinical utility of the test.
The Work Loss Data Institute's clinical practice guideline on "Low back - lumbar & thoracic (acute & chronic)" (2011a) listed FAD as one of the interventions/procedures that was considered, but was not recommended. Furthermore, the Work Loss Data Institute's clinical practice guideline on "Neck and upper back (acute & chronic)" (2011b) listed discography as one of the interventions/procedures that was considered, but was not recommended.