Chronic back pain – when research comes out of the blue

Something potentially amazing just happened. I’m not being flippant, a randomised controlled trial (RCT: still the only research method that can genuinely tell whether a treatment works) from China has just produced results in chronic back pain that can only be described as amazing. The temptation is to say “unbelievable”. This trial published in the worlds premier pain journal might have completely passed me by had it not been accompanied by an editorial with the headline “A cure for back pain?” written by Professor Nikolai Bogduk. If you think that headline is an attention grabber (although note the question mark), the results of the trial are even more so.

The trial compared injecting a substance called methylene-blue (which is known to kill nerve fibres) into the discs of chronic back pain sufferers with placebo injections. The idea is that in chronic back pain the degeneration that affects the intervertebral discs causes increased nerve growth and this sensitizes the disc. The brain produces pain in response to the increase in noxious input. Patients were selected if they had “discogenic” chronic back pain (i.e. back pain caused by a sick disc). This was established using discography, a diagnostic test during which the discs are injected to provoke pain.  If this provokes the patients back pain symptoms then the test is considered positive.

The results of the trial are show-stoppers. 73% improvement in pain and disability for the real therapy versus 1% and 3% improvement respectively in the placebo group. 91% patient satisfaction with the real treatment versus 14% for placebo. To give you an idea of scale there are no therapies out there for chronic back pain that one could honestly describe as being more than moderately or marginally effective for chronic back pain, in fact nearly all of our therapies dance around the threshold of what we might consider minimal clinical significance (often defined as around 15% improvement) (for example see here and here). It is important to note that nobody is suggesting that here is a treatment for all chronic back pain. This trial suggests that there is a subgroup of people with “discogenic” chronic back pain, that can be identified by discography and that these patients are likely to be helped by this treatment.

So then, job done, disc-related back pain cured? Happy Days?

There may be reasons to be cautious and interestingly they are not in the trial methodology itself. These results literally blow everything else out of the water – and that is the problem.  Prof Bogduk points out that the trial appears well designed and without fatal flaws and yet the results are unprecedented.  He wonders whether cultural factors may have influenced patient reporting, but if blinding of the patients and the researchers was maintained then this shouldn’t be an issue. The point he makes is that when results are so incongruous with the existing evidence base final judgement should be reserved until they pass the ultimate scientific hurdle: replication by independent groups in different patient populations.

There are other reasons why the results are so surprising. The evidence regarding the accuracy of discography as a diagnostic test is conflicting (see here and here). Of course an inaccurate diagnostic test should be expected to disadvantage a therapy in a trial, not inflate the benefits as you would recruit more inappropriate patients. There are other existing treatments that also aim to destroy the nerve supply of the disc (such as radiofrequency or thermal denervation). As the authors of the trial acknowledge these treatments have not performed well in RCTs to date, so what is different about methylene blue? Another possible cause for concern is that lack of any meaningful response in the group that received placebo injections. This lack of response may be due to the fact that the earliest point that patients were followed up was 6 months after the treatment and we know that even in the short term the placebo effect in clinical trials is quite unstable. In other trials of similar disc treatments we see the same pattern. One trial demonstrated a 17% improvement in pain scores at 6 months with placebo treatment (versus 37% with the real treatment) whilst another trial found essentially no change at all in either group. Nonetheless the lack of any placebo response to such an elaborate and invasive procedure is curious.

The patients recruited in this trial would fall under the diagnostic umbrella of “non-specific low back pain” (which really means “we don’t know the cause”). This umbrella accounts for about 85% of all back pain cases. When we look at the factors that influence the prognosis of this type of back pain we find little of use. There really is no measure of spinal pathology that is predictive of outcome (including discography and modern imaging technology), in fact commonly the only factors that seem to influence outcome are the severity of the patients pain and a number of psychosocial variables. This had led many to move away from a model of chronic back pain that has spinal damage at its core and to the development of biopsychosocial models that recognise that the pain and disability may be influenced by non-spinal factors. Indeed in response to evidence of abnormal pain processing in chronic pain patients some have speculated that chronic back pain is a problem of central nervous system pain processing (also see here and here). If the results of this trial accurately reflect the efficacy of methylene blue then this model will have taken a significant knock and as scientists we may need to reappraise.

Finally there is a phenomenon in research known as small study effects. Often early small studies exaggerate the effect size of treatments and the ensuing large studies produce more conservative results. This study is neither very small nor impressively large but the methods appear sound.  It is simply unfair to dismiss these findings and to be clear none of the above discussion aims to do that. But when research findings are so at odds with the existing body of evidence it is fundamental to scientific enquiry to consider them as interesting and encouraging but then to wait and see what follows. I think a follow up blog in 3 to 5 years time would be a good idea and I am not yet ready to send my chronic back pain patients for a shot of blue, nor to encourage them to see disc degeneration as the fundamental cause of their problem. But that could change…..

Original Article Abstract

ResearchBlogging.orgPeng B, Pang X, Wu Y, Zhao C, & Song X (2010). A randomized placebo-controlled trial of intradiscal methylene blue injection for the treatment of chronic discogenic low back pain. Pain, 149 (1), 124-9 PMID: 20167430

A preliminary report of clinical study revealed that chronic discogenic low back pain could be treated by intradiscal methylene blue (MB) injection. We investigated the effect of intradiscal MB injection for the treatment of chronic discogenic low back pain in a randomized placebo-controlled trial. We recruited 136 patients who were found potentially eligible after clinical examination and 72 became eligible after discography……

For full abstract and article go here


Bogduk N (2010). A cure for back pain? Pain, 149 (1), 7-8 PMID: 20129734


  1. I personally found this quite interesting both as a previous healthcare worker and as someone that suffers from crippling back pain. 2.5yrs ago I was in an auto accident that ripped my disc between L4-5, thus giving me ddd, and schroliosis & arthritis from T12-L4. I have been through the gamet of therapies and medications (excluding narcotics) with little to no relief. The pain has made me change every aspect in my life, yet I remain optimistic in hopes that there will one day be a treatment that can decrease the pain. Not everyone has chronic pain with depression…some of us have doctors that do not listen to the symptoms and seem unable to create a care plan, rather are stuck at the level of medical remedies that were available when they went through med school. I am glad to hear people around the world are not giving up on patients like me!

  2. It is always interesting when new research arises. What we thought we knew isn’t true anymore…and what we think we know may not be true tomorrow!

  3. Any news on this subject?


  4. Excellent blog. I am more interested about chronic back pain research suggestion for back pain cure. I love it.

  5. After reading your article it makes me realise its very early days with this hopeful cure so I will not get to excited yet. Thanks for leaving a comment on my blog.

  6. david cook says

    i have suffered chronic lower back pain since a mva,1983 ive had surgery to l4 l5 disc 1986,since that time my pain as progressively got worse,i am to the point of trying anything,so i look forward to anything that may give some relief, this trial certainly as my interest.

  7. Hi Snippets,

    Some really interesting points. I am no discography expert myself (maybe somebody out there is who can help?) but skimming the literature suggests disagreement regarding how sharp a test it is. In the Peng study report it appears that the injection may have been delivered at the time of discography or at least that’s how it reads to me. However all people in both groups will have had a positive discography and so the follow-up would just include 2 groups of people who have been told the same thing.

    I guess the reason I find it so surprising is that we’ve known that discs undergo changes for a long time but in the good longitudinal data it doesn’t appear to have much influence on the course of the problem.

    As for the eccentric loading in tendinopathy thing it is a very popular therapeutic movement at the moment but I’m not sure the quality data can be considered conclusive yet. Take this from a recent systematic review on Achilles tendinopathy management ( :
    “Eccentric exercises were noted to be equivalent to extracorporeal shockwave therapy (1 study) and superior to wait-and-see treatment (2 trials), traditional concentric exercise (2 of 3 trials), and night splints (1 study). Extracorporeal shockwave therapy was shown to be superior to a wait-and-see method in 1 study but not superior to placebo in another.”

    So conflicting evidence for the superiority of eccentric exercise over another exercise approach and equivocal effects with a therapy that appears not to outperform placebo.

    If the results of the Peng study are right it suggests that in these patients the chronic pain is being driven primarily from the bottom-up rather than the top-down. I guess I have got used to not thinking about the problem in that way. As for the central neuroplastic changes one might expect in a chronic pain patient – well we don’t know enough but perhaps a plastic system is more dynamic and adaptable than we might give it credit for. Of course all of this speculation is putting the cart well before the horse of independent replication, and I would agree with Prof Bogduk that we need to see that before doing too much reappraisal.

  8. This just blows my mind. It would be great to see the full-text article.

    I have never observed a discography which means I may be inaccurate in my initial thoughts. Are the people within this particular study being diagnosed with discogenic pain in a somewhat different manner than the normal discography procedure? Baogan Peng has been focused on this particular type of research since at least 2003. The following link might be the initial work leading to the current work being discussed on this site. I do not have access to full text of the article shared on the blog, but I will assume the methods in determining discogenic probably didn’t change. Peng described a 2 part process in defining discogenic. So, if the definition of how discogenic is determined is the same as the earlier work, is this definition standard across the world?

    I’m interested in the timing of the person learning of the results of the discography, the explanation of the findings and when the injection occurred after learning the discography results. I can literally visualize the “bad” disc. The reason I am interested in the timing is because we have literature indicating how how test results and the description of what was found can increase fear and anxiety. I’m thinking there could be potential differences in results if the findings were shared and an immediate injection occurred versus the findings were shared and then having to wait to be scheduled for the injection and then waiting for the injection after the procedure was scheduled.

    Baogan Peng also reports differences between degenerative discs due to aging versus painful discs versus normal discs: And if the differences aren’t notable via MRI, another reason for failed back surgeries and poor outcomes with surgical procedures:

    In an odd way, Peng’s work is almost paralleling what Peter Andersson is mentioning with some conditions of the Achilles tendon – the altered neovascular structural changes.

    So, two things come to my mind (as a physical therapist) – the tendon can be altered to be normalized with eccentric exercise and the subjective complaints can be substantially reduced. So what about the intervertebral disc? Is there some way to normalize the neovascular changes or prevent the neovascular changes? The other question – the biopsychosocial factors… someone having pain for more than 6 months – what’s the scoop with that aspect? How was there not a higher percentage of non-responders? The central changes and pathways being utilized for 6 months or more just went away magically?

    Quite thought provoking work…


  9. Thanks all for your comments.

    @ Bronnie,
    Even if the results are replicated it doesn’t kill off the biopsychococial model, it just indicates that there is a subgroup of patients with chronic back pain who when identified by discography can be more successfully treated than we have ever seen. In saying that a recognition of the non-biological would still represent a mature and holistic treatment approach.

    @ Mike. Thanks for the link. I saw that fascinating paper on the detrimental effect of discography. Striking as the effects are they don’t tell us whether the increased disc degeneration was related to an increase in symptoms or a change in the clinical picture. Given the poor relationship between disc state and outcome in other studies I wonder how concerning these findings are? I dunno?! In relation to your chosen approach to management it certainly sounds reasonable but unless it is shown to outperform other approaches in well controlled trials I guess we don’t know with confidence that it is a better solution.

    @ Peter, there seem to be lots of different injection based approaches to tendinopathy right now. Once again I await the reliable RCT data with interest. Uncontrolled clinical studies tend to be cheeky monkeys that overestimate effects. If anyone is a little sceptical about the need for RCT’s read this fantastic story from the career of Bill Silverman (something of a hero) here:
    Or this fabulous editorial recently in the spine journal on vertebroplasty here: (sorry for linking to pubmed but I find the sciencedirect links don’t seem to work later on).

    Thanks again all for your comments

  10. The evidence on discogram has progressed -not only is it not accurate to determine the level of discogenic pain it advances the degenerative process ( A better solution is to restore the mobility at the level (carefully) and slow progress daily weigh bearing activities -that will take care of those pesky extra nerves. See Explain Pain by Butler and Moseley for more about pacing.

  11. Peter Andersson says

    Hi folks!

    Intersesting, kind of same treatment as in chronic pain in Achillestendon, inject a substance that kills nerves and vessels. In Sweden Aethoxysklerol is used quite succsessfully according to the clinics that do it…
    Exentric load of the tendon for about 3 months works as well…No one knows why but it 6 out of 10 get well…

    Best wishes, Peter

  12. An excellent analysis of this rather startling study! Thanks – I too am somewhat unsettled by the study, and the points you make about duration of follow-up, and placebo are well made. It will be interesting to see the study replicated in other settings and see results over time. Until then, like you, I’ll be staying with the biopsychosocial model and a cognitive behavioural approach to managing pain.