I’m not a vegetarian. Perhaps that understates the issue, I am a fundamentalist carnivore and the idea of a barbeque without sausages and burgers is to me an unconscionable aberration; an affront to common decency. Just be honest and call it a salad party….and then don’t post my invitation.
Just as the missing meat at a barbeque is a matter of personal concern to me, so some very well regarded back pain researchers have just published a discussion paper in the European Spine Journal suggesting that something vital is being ignored in most back pain research – the back itself. They suggest that with the rise of the bio-psycho-social model of low back pain we all forgot the “bio” and that research is generally not focusing on improving our understanding of spinal pathology. They then point out some errors of reasoning that may in part be to blame for this.
Their key points are as follows:
- Just because possible spinal diagnoses do not seem to improve clinical outcomes does not mean that they are unimportant in understanding back pain.
- You don’t necessarily need a “gold standard” diagnostic test to investigate new diagnostic approaches.
- Just because some folk without back pain demonstrate pathological changes does not mean that such changes are universally unimportant to back pain.
- Clinical guidelines dismiss the ability to diagnose back pain based on relatively weak evidence
- Using diagnostic tests to facilitate research does not mean you are endorsing them for clinical practice
- Ignoring the biological component based on the current evidence base is fallacious
I have sympathy for much of their case. In a very human way research questions tend to fit into popular paradigms and it is a fair observation that in the label “non-specific low back pain” seems to have lead to a collective acceptance that since current diagnostic approaches seem unsatisfactory we shouldn’t look for new ones, or try to refine the existing ones. It does not necessarily follow that a correct diagnosis will lead to better clinical outcomes, and neither does it follow that just because our existing diagnostic models are not great there is no spinal diagnosis to be made – we might be missing something important in the spine.
However I do feel that in stating their case some relevant evidence was not mentioned or too readily dismissed. For instance the authors argue that “almost no quality literature exists” looking at specific pathology seen on MRI as a predictor of the course or development of back pain. But I can think of a few studies that are worth a mention. Borenstein and colleagues recruited a group of 67 asymptomatic subjects who were subsequently followed up 7 years later. At the initial MRI scan 31% of participants demonstrated an identifiable abnormality of the lumbar disc or spinal canal. From the 50 who responded no association was seen between the incidence of disc or spinal canal pathology and the development or duration of low back pain. Similarly Jarvik et al. recruited 148 folk who reported no back pain and performed MRI scans at baseline and at 3 year follow up. 131 of this group were followed up of whom 123 had repeat MRI scans. The presence and course of any back pain was monitored at 6 monthly intervals. Over the study period MRI findings were found to change little and no association was found between new low back pain, disc degeneration or endplate changes, annular tears or facet joint degeneration. Finally Eugene Carragee’s group recruited 100 patients with mild persistent LBP without disability. Subjects underwent lumbar MRI and those who were willing underwent lumbar discography. Subjects were followed at 6 monthly intervals over a 5 year period. Again no association was found between any feature of MRI or discography and clinical outcome.
Now you couldn’t call this a wealth of data but these studies are prospective and I don’t think that they represent “almost no” data. They each have their strengths and weaknesses but their findings are reasonably consistent. The main issue is probably their size and lack of power to detect a relationship but they at least indicate that if a relationship is there it is probably pretty weak.
Regardless the authors make an important point: forgetting the spine in back pain research is likely to reduce our understanding of spinal pain. The current absence of evidence may not necessarily be evidence of absence and ignoring possible peripheral drivers of pain from the spinal tissues is just as undesirable as clinging to an old-school Cartesian view of back pain. You can bring your salad, your bread rolls, your ketchup and your cheese but a hamburger without meat is not really a hamburger.
Note to Aussie readers and colleagues – that does not mean I am endorsing the inclusion of pineapple or beetroot in a burger – clearly that is just plain wrong.
Addendum
It’s worth being clear that the paper and this blogpost are principally about back pain research, not clinical practice. There is no shortage of folk focusing on structural spinal problems clinically, both real and imagined. My own personal view is that no treatment model yet has a strong basis for action in light of the existing evidence (for both diagnosis and efficacy).
About Neil
As well as writing for Body in Mind, Neil O’Connell 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
He is currently fighting his way through a PhD investigating chronic low back pain and cortically directed treatment approaches. He is particularly interested in low back pain, pain generally and the rigorous testing of treatments.
References:
Borenstein DG, O’Mara JW Jr, Boden SD, Lauerman WC, Jacobson A, Platenberg C, Schellinger D, & Wiesel SW (2001). The value of magnetic resonance imaging of the lumbar spine to predict low-back pain in asymptomatic subjects : a seven-year follow-up study. The Journal of bone and joint surgery. American volume, 83-A (9), 1306-11 PMID: 11568190
Carragee,E., Alamin,T., Miller,J., Carragee,J. (2005). Discographic, MRI and psychosocial determinants of low back pain disability and remission: a prospective study in subjects with benign persistent back pain The Spine Journal, 5 (1), 24-35 DOI: 10.1016/j.spinee.2004.05.250
Hancock MJ, Maher CG, Laslett M, Hay E, & Koes B (2011). Discussion paper: what happened to the ‘bio’ in the bio-psycho-social model of low back pain? European spine journal PMID: 21706216
Jarvik JG, Hollingworth W, Heagerty PJ, Haynor DR, Boyko EJ, & Deyo RA (2005). Three-year incidence of low back pain in an initially asymptomatic cohort: clinical and imaging risk factors. Spine, 30 (13) PMID: 15990670
Dear Neil,
Thanks for commenting on our paper. We have got quite a few favourable comments like yours so maybe the tide is turning and people are ready to consider the ‘bio’ component of biopsychosocial.
When we said “almost no quality literature exists” looking at specific pathology seen on MRI as a predictor of the course or development of back pain we meant to say two things: (i)there was not much research and (ii) what was there was often not of high quality. For example with the Borenstein paper they enrolled a convenience sample of 67 people, re-imaged 31 and got data on back pain course on 50. The information on back pain was obtained by asking people to remember how they fared over the preceding 7 years. I cannot remember what I had for lunch yesterday so I would be very bad at remembering the data Borenstein reported in that study; such as days off work, number of back pain episodes, duration of pain, the diagnosis participants got from a care provider.
My view is that we need larger and better studies to resolve the uncertainty in this area. At the moment many people “know” there is no association between imaging and subsequent LBP but unfortunately that knowledge is at best derived from reading weak studies. More often their knowledge has come from reading summary statements in guidelines and texts. You have to be very careful here because a statement that there is “no evidence” can be misinterpreted as “evidence of no association”.
PS I am with you on hamburgers. I never understood the beetroot and pineapple thing and I was born in Australia. That is clearly wrong (p<0.05).
Chris
[Reply]
Neil O'Connell Reply:
July 22nd, 2011 at 5:17 pm
Hi Chris,
Great to have your contribution on the blog and many thanks for taking the time to comment. My own interpretation of the evidence to date is that from cross sectional and prospective studies the influence of structural pathology current seems underwhelming but I take your point completely that better evidence is clearly required.
In the big study by Cheung it is hard to see which variables they controlled for in their analysis but age looks a big player to me. Also adding up disc degeneration at any level of the spine with a fairly loose definition of back pain and no detail of severity and duration really clouds the picture I think.
That study of methylene blue injection by Peng is still very curious, I have seen from correspondance in PAIN that a few folk find the results a little difficult to wholly believe. We’ll havw to wait for the replication studies!
As for the beetroot thing I think the question has been answered. Further studies are both unnecessary and arguable unethical. Viva scientific consensus!
[Reply]
Dear Neil
Good points as ever. Clearly the biopsychosocial model is just that. Everyone has their ‘bag’ and there has certainly been a wave of ‘psychosocial’ work that of course has great value. On a face to face basis with patients, looking for the drivers that influence the different dimensions of pain to create the overall brain response and conscious experience is fundamental. We know that beliefs about the pain, exercising, doctors, physios etc will impact upon behaviours and hence require addessing and perhaps challenging with good quality education. But, to ignore what is going on locally in the spine would be to miss a trick. We know that pain is not an accurate indicator of tissue damage, however the state of the tissues is important, especially to the brain.
Commonly we hear that in chronic cases ‘the pain is not there for any great purpose’ as opposed to acute pain when the warning is essential for survival. I disagree with this on the basis that things are not right. Usually there are issues with motor control, a dose of fear avoidance, often guarding and other protective measures and perhaps local cycles of inflammation (including neurogenic) that easily ignite the danger pathways somewhere along the neuroimmune axis. The brain must then respond. That is it’s job. Further to this, we know that not moving properly or indeed restricting movement will bring about tissue change. All in all, there are very good reasons to look at the tissue end but of course this is one end of the spectrum with other influences including the immune system, endocrine system and autonomics.
Biopsychosocial is often banded around. A bit like the word ‘stress’. Diluted and used in a way that does not reflect the true essence of what it means. The biopsychosocial model I think is the way to think about pain, but not cherry picking 2/3 of the integrated dimensions.
How about cherries on the BBQ?
[Reply]
I have often described the poor correlation between MRI results and back pain as one of “the most clearly established facts in all of musculoskeletal health care.” It’s become a bit of a talking point. Alas, I see now that I may have been overstating it a little! I was probably too eager to take the published conclusions at face value — confirmation bias strikes again.
Nevertheless, Neil’s makes an important point: “if a relationship is there it is probably pretty weak.” Amen. How much can the “bio” possibly matter in light of those results, imperfect as they are? If detectable pathology reliably causes pain, there should really be no ambiguity. Whatever the actual relationship is, weak or strong, I’m interested in it academically, and eager to see more evidence about it. By all means, let’s keep trying to find out exactly how much pain is caused by which structural pathologies.
But to justify the typical clinical approach to back pain — definitely still biomechanically minded! — the casual relationship has to be loud and clear, none of this iffy stuff. While the existing evidence is obviously not complete, and I was silly to ever think it was, it does still seem to put some pretty strict limits on the significance of the “bio” …
[Reply]
Hi all
Of course we can not conlcude that images matching or not matching clinical presentations because there isn’t firm evidence.
But to this imaging findings, we should add the interventions based on imaging findings or EMG (I mean surgery), and the succesful of those.
Bio is underestimated, in the way that it seems useless to make any clinical decision, at least that it matches with the clinical presentation.
Anyway, we should take the authors’ conclusion just as a call to keep in the way, and not to go away too quickly.
Good discussion.
[Reply]
Hi all,
Great to see some discussion of this issue as this was our primary aim for the paper.
As Chris mentioned our summary of the existing literature on the link between specific pathology seen on MRI and the course or development of back pain was “(i)there was not much research and (ii) what was there was often not of high quality.” Paul, Neil and Chris have discussed some of the low quality but negative studies. I want to point out that there are also low qulaity studies that do support a reationship between pathology and low back pain. The following example is copied from Ciaran Williams Honours dissertation last year.
“Salminen et al.(1995, 1999) studied thirty-eight school children that reported a history of recurrent low back pain and matched controls that reported no history of low back pain. MRIs taken at baseline when participants were 15 years old, were assessed for disc degeneration, disc space narrowing, disc protrusion, Scheuermann-type changes, central stenosis and muscular atrophy. At baseline and a 3-year, and 8-year follow-up, participants were questioned about low back pain ever (lifetime incidence), low back pain in the past 12months, and recurrent or continuous low back pain.
Baseline disc degeneration, disc protrusion and Scheuermann-type changes were all significantly associated with “recurrent or continuous” low back pain at the 3-year follow-up. For instance, 53% of those with baseline disc degeneration compared with 19% without disc degeneration reported “recurrent or continuous” low back pain at the 3-year follow-up (p=0.007). And 83% of those with and 23% of those without baseline disc protrusion, reported “recurrent or continuous” low back pain (p=0.002). Therefore, disc degeneration and disc protrusion may be prognostic factors for recurrent or continuous low back pain.”
My other broad point is that we know so much less about this area than most people realise. As a result we may be doing and interpreting these studies in ways that hide potentially important relationships. For example 1) is there an important thershold for degeneration (most studies dichotomise for analysis). In a paper we have in press the thershold made a major difference to the realtionship betwen degeneration and low back pain. 2) we really don’t know if combinations of findings are more important or different than single findings. 3) we lack reliable measures for many of the MRI findings 4) is patholgy relevant to some outcomes (eg recurrence) but not others (eg time to recovery).
Another interesting paper on this topic people might like to read is
Duncan R, Peat G, Thomas E, Hay EM, Croft P. Symptoms and radiographic osteoarthritis; not as discordant as they are made out to be? Ann Rheum Dis 2007 66(1): 86-91
Cheers
Mark
Salminen JJ, Erkintalo M, Laine M, Pentti J. Low back pain in the young. A prospective three-year follow-up study of subjects with and without low back pain. Spine. 1995 Oct 1;20(19):2101-7
Salminen JJ, Erkintalo MO, Pentti J, Oksanen A, Kormano MJ. Recurrent low back pain and early disc degeneration in the young. Spine. 1999 Jul 1;24(13):1316-21.
[Reply]
Great discussion all and many thanks particularly Mark for taking the time to discuss your paper here. Nice to be given some more bedtime reading to do. It will be fascinating to see how this strand of research pans out….
[Reply]
Dear Neil
I really like the hamburger analogy and your discussion is terrific. In Australia, as in other parts of the western world we are seeing soaring rates of obesity, linked to heart disease, diabetes and other chronic illness…perhaps adding to the burden of LBP?
I would like to pose the idea that researchers not forget the beetroot and pineapple in the hamburger, given it is always the tinned stuff. Lil
p.s. I’m the cheeseburger type, not a researcher & work voluntarily for APMA
[Reply]
GDay all -
this is a great conversation. Thanks Neil for yet another great post and thanks a million Chris and Mark for contributing – excellent stuff. I had a few thoughts on the paper too. On the basis of my observations, I agree with Richmond’s comment that the biomechanical, or I would say ‘structural/pathology’ paradigm is still the dominant approach to back pain – I don’t believe that this aspect has been forgotten by the clinical or research community. There are entire journals dedicated to using the structural/pathology paradigm to treat chronic pain and they do not seem to be short of submissions. Be that as it may, I also wholeheartedly agree with the notion that activity in primary nociceptors usually upregulates pain and to ignore this piece of the puzzle would seem folly.
I find the whole biopsychosocial label really misleading. That is, structural/pathology issues are not, in my view, the only piece of the puzzle to which we should be attaching the label of ‘bio’. Changes in the response profile of central nervous system cells, change in glial expression, changes in gray matter density, upregulation of endocrine, adrenergic and other mechanisms are all biological factors. Cognitive and contextual issues can only have their effect on pain via biological mechanisms. I think what we mean by biopsychosocial is actually more like anatomopsychosocial, in which case I am definitely an anatomopsychosocialist. I believe there is irrefutable evidence that primary nociceptor activity can modulate pain. I also believe there is irrefutable evidence that primary nociceptor activity is neither sufficient nor necessary for pain, so a tissue cannot, in my view, be a source of pain. I imagine that that does not surprise anyone. Keep up the chatter – it is refreshing reading I must say.
[Reply]
Thanks Lorimer for the wonderful comment.
[Reply]