Antibiotics for low back pain revisited. Important questions asked.

Back in March we blogged about a new and intriguing trial published in the European Spine Journal (ESJ) that appeared to demonstrate that for a group of low back pain sufferers, a course of antibiotics may be an effective treatment. We got to this trial early, when it was published online but not yet in print, and we reported it in a gently critical way.  Since then so much has happened.  The trial was published to great fanfare.  Press releases claimed “cure rates” for a substantial proportion of patients.  Now, the term “cure rate” in relation to low back pain is very bold indeed and it is bound to draw attention. And that’s before the term “worthy of a Nobel prize” was thrown into the mix.

With all the noise (hubris?) and substantial media coverage of the trial, came critical scrutiny and no small measure of sceptical discussion. Margaret McCartney wrote a great piece for the BMJ that raised some critical questions with regards to aspects of the trial (and kindly cited us!) and also pointed to possible conflicts of interest.  The doctor and blogger Benjamin Dean feistily expanded on this latter issue particularly, both in his blog(s) and in letters to the journal’s editor. Others also wrote letters raising various points.

But a new letter just published in the ESJ, the same journal that published the original papers, really caught my eye.  In his letter (infuriatingly for public discussion behind a steep paywall), Svend Lings raises 14 questions about the study, the data underlying its basic premise, the methods and results of the trial itself and the issue of possible conflicts of interest. It touches on some very interesting points that I had not begun to consider. For example, is it possible that the initial findings of infection were actually the result of contamination? Why does the study report appear to differ from the trial registry report and, if it is indeed true that infection entered the bloodstream initially from brushing the teeth, and assuming the treatment is itself effective, what happens the next time this happens? Another 100 day course of antibiotics? Is this viable at a population level?

Credit also to the journal editor for publishing it, though I would humbly suggest that this is a discussion that all should see, not just those with academic subscriptions or the means to pay.

Neil O’Connell

Neil O'ConnellAs well as writing for Body in Mind, Dr Neil O’Connell, (PhD, not MD) is a researcher in the Centre for Research in Rehabilitation, Brunel University, West London, UK. He divides his time between research and training new physiotherapists and previously worked extensively as a musculoskeletal physiotherapist.
He also tweets! @NeilOConnell
Neil’s main research interests are chronic low back pain and chronic pain more broadly with a focus on evidence based practice. He has conducted numerous systematic reviews including some for the Cochrane Collaboration. He also makes a mean Yorkshire pudding despite being a child of Essex.
Link to Neil’s published research here. Downloadable PDFs here.


Svend Lings (2013). Antibiotics for low back pain? European Spine Journal DOI: 10.1007/s00586-013-2977-0


  1. Dawn Hustler says:

    Hi there
    I had a DISCETOMY 17/4/13 as I had a large herniation at L5/S1 and debris.
    Although it improved things for me(don’t now need a wheelchair to get about outside, but every step is very painful and everyday I might as well write off the mornings as it’s really difficult.
    I have recently had an MRI and it showed I’ve got swelling there now both sides of spine and a cyst.
    I have been put on the 100 days of antibiotic, Augmentin 375 mg, 3 times daily.
    I started on Wednesday 11th, watch this space! Anyone else currently on the treatment?

  2. Cynthia Jones says:

    I suffered with lower back pain and severe sciatica for two years without any long-term relief either from pain management shots or physical therapy. I was referred to a surgeon who assured me that my pain was the result of a severely herniated disc and surgery at that point was warranted and my only option. At the same time, I happened to be diagnosed with a urinary tract infection and treated with Cipro. Within three days, the sciatica pain had resolved around 75% and was completely gone within a month. It hasn’t returned in over a year. Within three months, the strength in my leg returned to the level before the sciatic appeared. In my case there was an obvious correlation between the pain and an infection. I haven’t been back to the Institute honestly because I haven’t needed it. I do plan on making an appointment to assess the current condition of my back. I’d like to see what exercise I can add to what I’ve currently doing since I’ve been feeling so much better. I agree. This needs to be understood better so people in the same situation as me can find the type of relief I did. I lost two years of my life to the severe pain before being treated for the infection.

  3. Louise Sheppard says:

    Hi Neil (and other wise people!)

    I’m reasonably new to properly critically appraising research (being in the throes of a literature review for a masters at the moment) So, I decided to “practice” the process on this article – however, one of the major issues I had with the methodology doesn’t seem to have come up in any of the online discussions I have found to date. I’m wondering therefore, if I am “making mountains out of molehills” and hoped that you might be able to comment…

    Is it an issue that they included both people who had had surgery and those who hadn’t, but they failed to report either:

    a) the time since surgery (as far as I can deduce it may have been any time up to 6 months before the start of the trial)
    b) the number of “surgical” compared to “non-surgical” subjects in the active and control arms (to ensure that they were indeed the same at baseline)

    My sense is that, if you have a group full of people who have had surgery they might be on a very different road to recovery than a group full of people who hadn’t, as this may well have influenced the results….

    Thanks! Louise

    Neil O'Connell Reply:

    Hi Louise,

    I am not sure thats a particular biggie – the rate of previous surgery for disc herniation is presented in table 2 (51.9% in the abx group and 40.3 in the placebo group, non significant), though you are correct that the time since surgery is not reported.

    However if these proportions had been different between the groups then that might indeed be an important confounder.

  4. Michael Ward says:

    Thats interesting Dennis – but make sure you are investigated and treated properly. I am glad it seems to have worked in your case, but I also agree with Neil in that a trial and error approach to the use of antibiotics in all sufferers of chronic LBP is not viable. We need to establish clear criteria for its use.


  5. Hello Neil
    I am a 10yr plus sufferer of chronic low back pain. In 2008 I had a microdiscectomy for a very large extruded l4/5 disc segment which was causing foot drop and severe leg pain. I had very good result fro surgery. However my back pain returned and despite physio/chiro/Pilates/NSAIDs it remained. Then I read the original article and thought what the heck I will try it. So I took 1/2 an augmentin duo forte every day and immediately my pain improved by 80%, I stopped after 60 days and backlit came. Subsequently I restarted and it has largely resolved once again. For me the proof is there, it may not work universally but it’s worth a try.

  6. Michael Ward says:

    Thx Neil for the reply. I fully agree that any treatment should be supported by an accurate diagnosis.
    re “widely treated” I wasnt aware that there was the suggestion that it would be appropriate for the wider population. I understood that it would be for a select group of patients: a specific pain profile, with imaging (modic) changes. Of course it would be nice to have an accurate antibiotic sensitivity but in the absence of this….
    re “crisis of population” – I am not sure what the percentage of chronic pain patients that this would apply to (and then what % of the population that would be), but I also understood that the greater threat of multi-resistant organisms was the indiscriminate use of antibiotics in animals (such as a growth promotion) with the risk of the MR bacteria entering the food chain or alternatively frequent use of uncompleted courses of antibiotics (eg URTI). Would this argument be used for cystic fibrosis? Broad spectrum antibiotics in a group with high risk for re-infection?
    Re “re-infection given the pathway”:re-infection rate is highly speculative (is there evidence for other discitis reinfection, or septic arthritis?) But in any case, is the suggestion that we withhold antibiotics if they have another form of discitis?
    Bottom line to me is: it is a small subgroup that this intervention might apply to; cost is not the question at this stage of research; multi-resistance is a low order problem in a small well supervised group; apart from the obvious ethical issue that if you know that someone has a condition that is responsive to a specific intervention, do you withhold it because of the risk of its recurrence? Hence my reference to cancer, or cystic fibrosis?
    Thx and I do appreciate the critical analyses.

  7. John Barbis says:

    Dear Neal,

    I appreciate your keen analytic sense and willingness to ask questions. I appreciate your willingness to be a gadfly, butI must admit that at time I get the sense that if you were given a glass of extremely fine wine, you would find something wrong with it: “The glass isn’t quite right to demonstrate the color; The air temp is not quite right to let bouquet reach the right part of the nose; etc” After reading your work, I feel that somehow I have wasted 35 years of effort and stollen from patients and insurers because nothing that I have done clinically works.

    I agree that there are some concerns about these studies, but I do not know of any study where that can not be said. I think that the effort and care that went into these two studies is truly commendable. Perfect no. Are there things to learn? Yes. But the effort and expense of the processes involved in gaining this information is mind blowing to me. I think that yes this is not the definitive study but, at the same time, I think it is a hell of a good start in exploring an area that others have thought about and done nothing with.
    It is standard protocol with many joint replacement specialists to make sure that the dentition and oral cavity is well evaluated and problems fixed before surgery. I am facing bilateral knee replacements next month and have been meticulous about oral hygiene. My dentist reminded me yesterday, to make sure I take my antibiotics before I see him in the future after the replacements. The model for bacterial infection of joint replacements is accepted. It makes sense to me that the modic change could be bacterial in origin. I applaud this group for attempting the difficult process of proving that.

    One final note about financial conflicts. I agree their financial conflict should have been mentioned, but I do not know a researcher who does not have a financial conflict with any of the work they do. Every good researcher that I have been in contact with makes money off their research in the lecture circuit, payment of travel expenses to conferences, and through the use of research for promotion. Because a financial conflict was not mentioned (it should have) does not negate the results of their research any more than the results of L’s research because he accepts ( and rightfully so) honoraria for travel or speaking ( not mentioned in his published research). Any one who thinks just mentioning a researchers financial interest removes conflict is being extremely simplistic.
    I totally agree that media have misplayed this study. There are concerns. If all initially positive studies into AIDS, various cancers, etc. had turned out to be true, we would be living in a much healthier world. But even if the results are not pristine, there are important leads and concepts demonstrated that deserve serious study. In addition I, as a clinician, have to seriously consider the “half full glass” component of this study in my potential thought process on management. JOhnB

    Neil O'Connell Reply:

    Hi John,

    I had to google “gadfly”! I take your point, to a point. I certainly do err on the more critical side of life, though I would not agree that we are being overly harsh in this instance. At no point have I dismissed the findings of this trial but the art of critical appraisal is to be aware of all the reasons to be circumspect, as well as the ones for optimism. Taken another way we have a duty to seek out all the space for the null hypothesis to be true – because only by looking for it and not finding it can we be more certain we are on the right track. I agree that a conflict of interest does not negate a study’s findings; it just creates more cause for circumspection. I agree that there is plausibility in the model presented. With this particular topic you are quite correct that this may indeed by a remarkable achievement and discovery, but given the history of false dawns we should be measured and await confirmation. I would suggest that I have been measured in this case, but of course it is always a fine line and one man’s critical appraiser is another man’s wrecking ball.

    In the broader sense, in my anecdotal experience the world of rehabilitation is often insufficiently self-critical, both clinically and in research. I myself have practised clinically for just shy of 20 years, and while I am sure I have made a positive difference in that role I am less compelled by the specific efficacy of many classic components of my “craft” and what I have learned in my capacity as a researcher has offered a more structured way of looking at those doubts and concerns. Being critical is often uncomfortable, but I am very comfortable with that!

    David Colquhoun Reply:

    I’m rather horrified by you statement that

    “Every good researcher that I have been in contact with makes money off their research in the lecture circuit, payment of travel expenses to conferences, and through the use of research for promotion.”

    That certainly isn’t true in the sort of basic research that I do. In the UK essentially all talks are given for nothing but expenses (and they aren’t income). Sometimes you get a modest honorarium in the USA, but that’s about it. I’m puzzled about what you mean by “the use of research for promotion”? Does that mean that companies pay you to promote particular treatments?

    If that really is true in clinical areas, it’s a serious concern.

    John Barbis Reply:

    Thanks for your response Neil ( sorry about the Neal part). Your points are well taken and I agree that we do not have sufficient gadflies ( I was going to use the term curmudgeon, but decided against it. I have always liked the sound of that word) in rehabilitation. I, often feel as a clinician, who tries to be evidenced based and artful in my care, that I am placed between a rock and a hard place in trying to process my understanding of the literature ( and lack of it) and the need to deal with the hurting patient who is sitting in front of me. The science and art of health care can be at once illuminating and daunting as we look at what we do and we try to find that pathway which is hopefully best for the patient. It is hard at times when we look at the hard science of research not to throw out the baby with the bathwater. Thanks for your efforts Neil.

    As far as seeing being horrified as seeing research as a business and a source of promotion, income, etc., come on. Have you ever tried to put on a major conference and bring the best researchers? Do you know the costs of bringing in high quality people? In most situation, high quality is equivalent with the best known researchers. At least in the US, it is not cheap. What researcher, especially in academia, does not see their research (not in toto, but at least in part) as a means to academic promotion, potentially new jobs, grants for future research, potential products, etc.? And in fact- There is nothing wrong with or bad about that. I have no problem with a researcher who has done high quality and valuable work being paid for it and honored for it not just with verbal praise but with hard financial outcome. My professional life depends on their work. We need a more mature understanding of research. Where funding is involved, I think Neil is correct, we need to be circumspect about the results. BUT FOR GOODNESS SAKE, LET’S PAY RESEARCHERS WELL. OUR LIVELIHOODS DEPEND ON THEIR WORK. Johnb

  8. Michael Ward says:

    “f it is indeed true that infection entered the bloodstream initially from brushing the teeth, and assuming the treatment is itself effective, what happens the next time this happens? Another 100 day course of antibiotics? Is this viable at a population level?”
    Isnt this an over-reaching critique? What if we said – what happens next time the breast cancer / leukemia comes out of remission – another course of very expensive chemotherapy? is this viable on a population level?

    I agree that the peer review is of vital importance – and we rely heavily on smart people like yourselves to review from an expert perspective, but to me it has sign-post significance: When I see a patient with a raging disciitis (probably staph) or a septic hip (possibly from a minor scratch on the leg) and see the response, it does not seem so unreasonable that there is a spectrum of disc infections, some of an almost occult presentation, many from unclear pathways. Also, I dont know where the suggestion comes from that there will be repeated infections in the same person.

    In any case, isnt this a very positive perspective that there will be a proportion of patients (that will require further confirmation and clarification of diagnostic criteria) that will respond to a very simple strategy. Given other options such as 3 months in a pain clinic being given strategies to cope with pain or a fusion, I know what I would prefer.

    Neil O'Connell Reply:

    Hi Michael,

    Thanks for your comment. I think the issue of repeat infection is interesting as the proposed pathway would appear quite prone to repetition. But that is if the infection is real in the first place. To be clear I am not trashing the study and neither is Dr Lings but there are serious questions to be clarified, either by better reporting or by more research.

    The population level issue is quite specific to antibiotics though. It seems we may be approaching a crisis of population level antibiotic resistance, largely due to overuse. So any condition that might be widely treated with long-duration and high dose antibiotics is a potential concern. Thats not to say we shouldn’t do it, but we need to be pretty certain those benefits are real, because there is good reason to believe the risks are.

  9. Thanks Neil for the heads-up on the very interesting letter which indeed poses even more questions about the original studies.

    Just to throw something else into the mix about bacteria and pain, there is new research by Clifford Woolf’s team published in Nature [Bacteria activate sensory neurons that modulate pain and inflammation. Chiu IM, et al Nature. 2013 Aug 21], which shows that Staph. aureus bacteria directly activate peripheral pain-sensing neurons, independent of local inflammation and the immune response. Furthermore, the bacteria-activated nociceptors suppress innate immune activity during acute infection.

    Unfortunately this article also sits behind a paywall, but you can freely access an excellent review of it on the Pain Research Forum website [].

    More bugs for thought!


    Neil O'Connell Reply:

    Thanks Julia,

    I will take a read. A nasty but nifty idea that. That bugs can tinkle the noiceptive ivories on their own.

  10. Nick Henschke says:

    Hi Neil,

    Thanks for an interesting article and the link to this letter! I have been following the reactions to this paper since it came out with some interest.

    A number of these points (especially regarding the reported data in the tables and confusion over the methods) seem to me to be a result of a less than rigorous peer-review process. Perhaps they were mentioned during the initial peer-review and responded to appropriately by the authors. Perhaps the reviewers did not even notice the irregularities. Perhaps they were too busy to give quality peer-reviewing the time and attention it requires!

    I think It might be helpful if more journals made available the pre-publication history (reviewer comments and author responses) of published articles.


    Neil O'Connell Reply:

    Thanks Nick,

    I totally agree – the BMC model is interesting in that respect – the peer review process is all there for those who care to look, at least for the submission to that journal. Imagine a system where all of the peer review process for every journal a paper had been submitted were published in full. Now that would be revolutionary.

    The more peer-reviewing I do the more I realise that it really is a mixed bag. So many times I ‘ll write a detailed and long review of a paper picking up what I think are a list of issues that, at the very least need acknowledging in the discussion, only to find that my co-reviewers have essentially waived it through. This seems worse in some fields than others.

    I guess in part it is because this is a human activity and the old positivity biases are clear and present. Also I think in some fields there is a tendency for papers to be reviewed by folk who are already invested in the idea that a treatment or theory has worth – and are excited enough to see more confirmatory evidence that they are tolerant or inattentive to basic methodological issues or grey areas.

    Given the topic and findings of this trial, the potential implications on a number of levels, we might have hoped that the review process ensured a little more clarity.