In 1988, Modic et al. suggested a classification scheme for endplate changes associated with erosive degenerative disk disease (EDDD) at the lumbar spine. Type 1 is characterized on a magnetic resonance imaging (MRI) scan by low signal on T1 and increased signal on T2 images, indicating bony edema 1 . EDDD is mostly a French denomination for a type of non-infectious erosive disease of the spine, among others, such as crystal deposition disease, erosive discopathy in patients under chronic hemodialyses and rheumatic diseases with inflammatory discopathy. EDDD is the degenerative, or osteoarthritic, subcategory of noninfectious erosive disk diseases. Erosions are not an essential feature of Modic type 1 changes, but the two have often been associated, since studies have described a constant presence of Modic type 1 changes in EDDD 2 . Moreover, usually superficial erosions have been described as one of the elementary bone lesions in degenerative disk disease associated with Modic type 1 changes 3 . Since ‘Modic type 1 changes’ is supposedly a radiological definition and does not encompass the clinical aspect, it has been accepted for a long time, in France at least, that we speak of EDDD, recently changed to ‘active discopathy’.

The correlation between Modic type 1 changes and back pain is still a matter of debate. There is however, to this date, far more evidence in favor of a real relationship between the two than otherwise. A 2008 systematic review 4 thus found ‘a positive association between vertebral endplate signal changes (VESC, or Modic changes), and low back pain (LBP) in the majority of studies reporting on this subject’, with Odds Ratios in the studies that reported a statistically significant positive association ranged between 2.0 and 19.9. Moreover, ‘a positive association between VESC and LBP has not only been found in the majority of studies of patients with LBP from different countries, but also in the general and working populations. In other words, there is a considerable consistency in this association’ 4 .

Another 2008 review 5 confirmed the possibility of asymptomatic individuals with Modic changes but emphasized on its infrequency 6 7 8 , and concluded that type 1 changes are ‘likely to be inflammatory in origin and seem to be strongly associated with active low back symptoms and segmental instability, thus reflecting a state of active degeneration and biomechanical instability of the lumbar spine’. For an MRI classification, Modic changes are therefore usually considered a validated clinical finding; Toyone et al. observed that 73% of patients with type 1 changes had low back pain, making asymptomatic subjects much less common than those with type 2 changes (11%) 6 . Modic type 1 changes have also been found to be specifically associated with the presence of low back pain 9 , supporting the hypothesis that this early stage of the Modic disease is active 10 . Moreover, patients with Modic type 1 changes usually experience acute flares of previous common chronic lower back pain, usually with an inflammatory pattern 11 . Therefore, the clinical presentation of EDDDs mimics that of spondyloarthropathies (SpA), with consequent first-line treatment based on the use of nonsteroidal antiinflammatory drugs (NSAIDs) 12 . However, these treatments can be inadequate or even totally ineffective (at least 25% of patients are said to be refractory to NSAIDs), and sometimes poorly tolerated (gastroduodenal ulcers, arterial hypertension, and so forth) 13 . Corticosteroid therapy, which is not indicated for common low back pain, has been suggested for EDDDs by Maigne and Ballard, but over 50% of patients in the only open-label study experienced therapeutic failure 14 .

Local treatments have been suggested, ranging from intradiscal cortisone injections to arthrodesis 15 16 17 18 19 20 . The results are, however, controversial 21 22 . These invasive procedures expose the patient to a risk of infection for intradiscal injections and to operating complications for arthrodesis 19 .

Pamidronate, intravenous bisphosphonate with action on bone turnover, has shown some efficacy in open-label studies on spondyloarthropathies that are resistant to NSAIDs, whether on ankylosing spondylitis or SAPHO (Synovitis, Acne, Pustulosis, Hyperostosis, Osteitis) syndrome 22 23 24 25 26 27 28 29 30 31 . Although this efficacy in spondyloarthropathies remains a topic of discussion, it has nevertheless been clearly demonstrated for various types of bone pain (Paget’s disease 32 , fibrous dysplasia of bone 33 and vertebral compression 34 35 ). Therefore, we conducted an open-label study on the efficacy of pamidronate in ten patients with EDDD in 2006, with conclusive results 36 . At one year, four patients deemed their overall improvement to be excellent (85 to 100%), four to be good (70 to 85%) and two to be poor (< 55%). The inflammatory pain pattern disappeared as early as the first month in four patients and after one year in nine out of ten patients. Mean scores on the Visual Analogue Scale (VAS) showed gradual improvement, which was significant at six months (35.1 ± 30.2 mm; p = 0.05) and one year (32.1 ± 24 mm; p = 0.01). The improvement was quickly significant on the Quebec scale (a 20-item self-administered instrument designed to assess the level of functional disability in individuals with back pain), at one month (26 ± 27.7; p = 0.02) and at one year (15.8 ± 18; p = 0.001). The mean Schober score, measuring lumbar flexion mobility, was unchanged.

The objective of this clinical trial is to demonstrate the efficacy of pamidronate in degenerative disc diseases by conducting a prospective, randomized, placebo-controlled study. Patients with inflammatory low back pain and a Modic type 1 change on their MRI, whose inclusion and exclusion criteria are detailed in Table  1, will be randomly assigned to one of two groups: one group will be given pamidronate and the other placebo. Pamidronate will be administered at a dose of 90 mg per day for two days consecutively. The primary outcome measure is a between-group difference of 30 points on a 100 mm VAS at three months. Secondary outcome measures are improvement in functional status and the drug’s safety. In case of failure at three months, it will be recommended that the patient, irrespective of the treatment initially received, wear a rigid or semi-rigid dorsolumbar brace, whose efficacy will be evaluated using the same criteria at six months to ensure the patient is at least partially relieved at the end of the study, provided that they had not used a brace before the start of the study.

MRI, magnetic resonance imaging; NSAIDs, non-steroidal anti-inflammatory drugs; VAS, Visual Analogic Scale.

The treatment options for Modic type 1 disc diseases remain quite limited; invasive local treatments, such as intradiscal injections of cortisone derivatives or spinal arthrodesis, have generated controversial results 21 22 , and the brace remains a last-resort solution for patients who wish to maintain some mobility and often consider it cumbersome and unsightly.

There are, to date, no published controlled trials involving pamidronate in EDDD. The results from our pilot study (an open-label study with ten subjects) were encouraging 36 , and therefore we hypothesized, in light of this pilot study and the widespread use of pamidronate in painful bone diseases, that pamidronate could be effective in relieving EDDD-induced pain. However, the efficacy of pamidronate in this condition clearly required further investigation in a prospective, blinded, randomized controlled trial. Subject recruitment began in January 2013 and will continue until the end of December 2014, with the results made available in March 2015.

Regarding safety concerns, vitamin D blood levels will be measured before the product is administered and patients will be given vitamin D supplementation accordingly prior to treatment 44 . Every patient will also be examined by the dentist affiliated with the study to look for, and if necessary treat, contraindications to intravenous bisphosphonates.

All studies include biases, and we have taken several steps to limit those that could arise in this study; as pamidronate can cause fever and flu-like syndromes, the patient will be included by one physician and evaluated by another. In addition, to minimize this effect, patients from both groups will receive 1 g of paracetamol prior to the infusion and 1 g of paracetamol every 8 hours for 48 hours. To guarantee the double-blind and prevent biases in monitoring and measurement, physicians from the Rheumatology Department will be responsible for administering the product and monitoring the side effects (such as hyperthermia in cases of patients receiving pamidronate despite the administration of paracetamol, which would result in unblinding for both the patient and physician), and physicians, physiotherapists and occupational therapists from the Physical Medicine and Rehabilitation Department will be responsible for assessments (D0, D45, D90 and D180). In addition, and again to guarantee the blinding for the patient, the products administered intravenously will be identified not by the product name but by the patient’s randomization number, and will be prepared by the hospital pharmacy so that the nurses performing the infusions will not know which product is infused to the patient. Treatment changes over the six months of follow up will be recorded and, to prevent a recall bias, each patient will be given a diary including a table to be completed following each change in his or her standard treatment. Finally, an experienced radiologist affiliated with the study and specializing in osteoarticular diseases is responsible for confirming each radiological diagnosis of Modic type 1 disc disease on an MRI scan. Concerning the sample size estimation, the assumption of our protocol is that the standard-deviation of the difference (test versus control) in change scores (initial versus final) is the same as the previous standard-deviation of the change score found for treated patients in the open pilot study 36 . Therefore, sample size estimation is based on previous works, which could be considered heterogeneous for the standard-deviation of the change score. In our pilot study 36 , the standard-deviation of VAS was 19.1 at baseline, 22.7 at three months and 24.1 for change score. As standard-deviation equals 30, our estimation appears near to the expected results and overestimates the value obtained in the pilot study. Nevertheless, blinded studies have a smaller effect which may not be captured by the standard-deviation. Moreover, the statistical power was voluntarily more important than 80% (1-β fixed equals 90%). Concerning the statistical approach, the primary outcome analysis will be pain-VAS difference at three months using a Student’s t-test if normally distributed and a Mann–Whitney test if not. In the secondary outcome analysis, a multivariate situation should be proposed using a linear regression model to take adjustment factors into account. Adjustments may threaten the randomized nature of the study. To further ensure the randomization of the study, a random-effect model, considering block as a random effect could be proposed.

PEPTIDE is ongoing and currently recruiting.

ANOVA: ANalysis Of Variance; CBC: complete blood count; CTX: C-terminal telopeptide; EDDD: erosive degenerative disc disease; EIFEL: French adaptation of the Roland-Morris Low Back Pain Questionnaire; ESR: erythrocyte sedimentation rate; FABQ: Fear Avoidance Beliefs Questionnaire; hs-CRP: high-sensitivity C-reactive protein; LBP: low back pain; MacTar: McMaster Toronto Arthritis Patient Preference Disability Questionnaire; MCII/PASS: Minimum Clinically Important Improvement/Patient Acceptable Symptom State; MRI: magnetic resonance imaging; MS: morning stiffness; NSAIDs: nonsteroidal antiinflammatory drugs; PEPTIDE: short for the French title “Etude Prospective sur l’Efficacité et la tolérance du PamidronaTe dans les dIscopathies Degeneratives Erosives”; PTH: parathyroid hormone; SAPHO: Synovitis Acne Pustulosis Hyperostosis Osteitis; VAS: visual analogue scale; VESC: vertebral endplate signal changes.

The authors declare that they have no conflicts of interest.

MS: conception and design, data collection, analysis and interpretation, manuscript writing, critical revision and final approval of the manuscript. SC: conception and design, data collection, analysis and interpretation, manuscript writing, critical revision and final approval of the manuscript. DA: design, data collection and analysis, critical revision and final approval of the manuscript. BP: conception, data analysis, manuscript writing, critical revision and final approval of the manuscript. CDa: data collection, critical revision and final approval of the manuscript. MV: data collection, critical revision and final approval of the manuscript. EC: design, data collection, critical revision and final approval of the manuscript. SM: conception and design, data collection, critical revision and final approval of the manuscript. SMG: data collection, critical revision and final approval of the manuscript. AT: data collection, critical revision and final approval of the manuscript. MC: data collection, critical revision and final approval of the manuscript. JJD: data collection, critical revision and final approval of the manuscript. AR: data analysis, critical revision and final approval of the manuscript. CDe: data collection, critical revision and final approval of the manuscript. All authors read and approved the final manuscript.