Brace yourself

Despite the popularity of exercise programmes that emphasise “core stability” for the prevention and treatment of chronic low back pain (LBP), numerous high quality randomised controlled trials (RCT) (e.g. here, here and here) have not found them to be superior to other therapies such as general exercise. In fact, core stability exercise programs have been found to be only marginally superior to placebo (e.g. here), and produce only small, short-term improvements in pain and disability.

For this reason, a recent study (Aleksiev 2014) caught our attention. It compared four groups of people with recurrent LBP who were allocated to receive strengthening or flexibility exercises, with and without abdominal bracing. It found that abdominal bracing added significantly to the effect of strengthening and flexibility exercises and concluded that “abdominal bracing could be considered as a preliminary muscle back belt” which “increases the frequency of trunk muscle contractions which increases the chance of successful LBP prevention.” Given that these findings are in marked contrast to the aforementioned RCTs, we wanted to see if there were any reasons for such markedly different results. Several noteworthy aspects to the study were addressed in our letter to the editor (O’Keeffe et al. 2014). These included:

  • A tremendously successful follow-up in this study, with not one of 600 participants lost to follow-up after 10 years. This is certainly not common – in fact, it is utterly remarkable!
  • The frequency with which exercises were performed was associated with positive results over the follow-up period, and this was much higher among those who were in the bracing groups. Despite this, the author has concluded that it was the abdominal bracing itself which caused the significant difference between groups. The author suggested that “Bracing probably reminded and convinced the patients to exercise more often”. In our opinion it is difficult to extrapolate this conclusion from the data provided.
  • Several CONSORT requirements were not fully detailed, such as baseline demographic and clinical characteristics for each group. Similarly, little detail regarding the interventions was provided, such as the length of the intervention, and how compliance was monitored. For example, it is hard to know from the paper how well participants continued to complete their exercises for the duration of follow-up.
  • No measure of disability was used.
  • No earlier results e.g. at 1, 3 or 5 year follow-up were described.
  • No detail provided on whether or not a blinded assessor was involved.
  • No measure to assess changes in trunk activation and co-contraction, making it difficult to justify conclusions about the mechanism of effect.

Due to the number of studies that contradict these findings, and the concerns we have raised above regarding shortcomings in the study methodology, we believe that the results must be replicated in other settings using more rigorous methodology before they become accepted clinical practice.

It is worth remembering that no relationship has been found between the onset of trunk muscle activity and pain/disability levels, or the degree of change in pain/disability after core stability training as previously discussed on BiM (e.g. here). Emerging evidence (see here and here) also reveals that disabling chronic LBP may be more closely associated with increased co-contraction and hyperactivity of “core” trunk muscles, guarded spinal movement and an inability of the spinal muscles to relax (including transversus abdominis and lumbar multifidus). Furthermore, increased focus on protecting the spine in a manner which is not seen amongst people without chronic LBP has the potential to increase hypervigilance, and reinforce pain-related protective behaviours that are known to fuel the vicious cycle of pain. There is now considerable evidence that LBP is a multidimensional disorder associated with a complex combination of physical, lifestyle, psychological, cognitive, social and neuro-physiological factors. Therefore, attempts to manage LBP are likely to require consideration of more than just the training of a specific muscle, or set of muscles.

About Mary O’Keeffe

Mary O'KeeffeMary O’Keeffe is a PhD student at the University of Limerick (UL), Ireland. The broad area of her research is the role of multidimensional rehabilitation in chronic low back pain. Her PhD research is examining whether tailoring this rehabilitation to the individual patient presentation enhances effectiveness, and is worth the additional time (and costs!) involved. Her supervisors are Dr. Kieran O’Sullivan and Dr. Norelee Kennedy from UL and Prof. Peter O’Sullivan from Curtin University, Perth. Their research group promotes evidence-based assessment and management of chronic pain through


Aleksiev AR (2014). Ten-year follow-up of strengthening versus flexibility exercises with or without abdominal bracing in recurrent low back pain. Spine, 39 (13), 997-1003 PMID: 24732860

O’Keeffe, M., Nolan, D., O’Sullivan, P., Dankaerts, W., Fersum, K., O’Sullivan, K. (2014) Ten-year follow-up of strengthening versus flexibility exercises with or without abdominal bracing in recurrent low back pain: Letter to the Editor. Spine, In Press.


  1. John Ware, PT says

    It’s inconceivable to me that this guy didn’t have some help from an office staffer, resident, colleague, his dog, -someone!- in devising the logistics or reviewing the manuscript. No one could put together a long-term trial of this magnitude without a lot of help.

    But, I see Michael’s point. Maybe those individuals decided they didn’t want to be acknowledged once the saw the final product.

  2. Mary: I would agree that this study does seem to have many issues that limit its acceptance at face value. Conceptually, core strengthening means many things to many people and to just accept that all so core strengthening is valuable or globally indicated is unwarranted. Certainly, approaches that detail activation of just one muscle, I think must be viewed with a critical eye as there is certainly no human movement or posture that is the result of a single muscle activation. The question of muscular strength, activation, control and appropriate relaxation probably are different across the spectrum of patients , making adequate control an challenge for any study. Thanks for your interesting post.

  3. Hi Jon.
    Yes, the language used is a tricky one, and it is not always clear apples = apples in some studies. but even that systematic review (Bystrom et al)has been evaluated as having some issues.
    “The authors’ conclusions appear too strong given the limitations of the data”

  4. Thanks for the post Mary. There’s significant nomenclature issues in the motor control literature but perhaps this systematic review in Spine on specific muscle activation (e.g. TrA, LM, pelvic floor leading to function) is relevant? Type it into Google and you can get full text via Research Gate

    Bystrom et al 2013 – Motor control exercises reduces pain and disability in chronic and recurrent low back pain: a meta-analysis

  5. Thanks Greg.
    The intention one I would be very interested in also so when you think of it send it my way

  6. Frédéric Wellens says

    I am astounded this core stability mirage is still an unsettled matter. I guess the proponents of this core stability theory will always find something to say to try and justify their ideas, the same way homeopathic medecine is still around.

    Kieran Reply:

    Agree Frederic
    I think it is hard to shake this one off, as it fits societal (and healthcare) views about the critical importance of spinal tissue structures over everything else

  7. Is the “Core Stability” being activated in a task or any activity of daily living. There is a big difference between whether the core stability is being activated so that the strength can be controlled and useful in ADL’s. Otherwise the exercise is similar to general exercise.
    I am not sure whether these studies have investigated this. A way to determine this would be another group that is instructed first/only to “push their feet into the ground” (rather than brace the core – the hypothesis being that the core will be automatically activated in this and all tasks of ADL

    Greg Holdaway Reply:

    The integrated activation of the ‘core’ along with other muscles that contribute to stability can be experienced with the decision to resist a gentle shove to the side. Actually, in accordance with published studies, the *intention* to resist is enough for the neuro-musculature to respond in an integrated fashion. The issue with postural stability is often (though not always) a lack of resistance to gravitational pull, with consequence distortion of the coordination between head, thorax and pelvis… an artificial tensing of the abdomen will temporarily ‘brace’ this relationship, but is not effective for ongoing postural support in and of itself.

    Wally Johnston Reply:

    I could not agree Greg. I am not aware that this has actually been studied ?

    Greg Holdaway Reply:

    Hi Wally,

    I was thinking of studies such as: Jeka JJ, Lackner JR. Fingertip contact influences human postural control. Exp Brain Res. 1994;100:495-502.

    The specific article showing intention to touch also has this effect escapes me a present…

    Wally Johnston Reply:

    Mean’t to say I could NOT agree more

    Kieran Reply:

    Hi Wally and Greg
    i don’t think there is any doubt that different exercises can have different effects on the activity of muscles (or muscle groups). the more debatable part is whether different exercises have a different effect on pain. so far, most studies (with their limitations of course) suggest any one type of exercise is only as effective as another type of exercise for pain

  8. What is Spine’s agenda that it passes this as EBM when it is missing so much? We are nowhere near April…

  9. Shovel of salt made me laugh out loud Ainsley R! Now, how do I go about employing whoever did their follow-up….

  10. Good to see ‘multidimensional rehabilitation” being studied for LBP. The conception, even just at the ‘physical level’, of specific exercise for specific muscle strength for something as complex as postural stabilisation plainly does not address the realities of the integrated sensory-motor interactions that are responsible for the generation and maintenance of trunk stabilising tone. Even a strengthened muscle can be under-utilised, the question of reflex neural activation of stabilising tone must be addressed if a consistent and effective approach to LBP rehab is to be developed.

    My particular bias in this is supported by 25 years clinical practise in postural movement re-education for both rehab and performance improvement. “Core Stability” sells well, but carries with it mis-conceptions of human functioning that can lead to further interference with normal postural homeostasis and respiration.

  11. I think I’ve seen some of those chronic pain patients who have trained to “abnormal brace”. What a mess to un-train….

    I’ve heard Paul Hodges say that he does not use terms like “bracing” or “stabilization” because they infer static holding and muscle patterns that looks nothing like normal core muscle control.

    Stu McGill does teach side planks for 70 seconds but never just that.

    Studies like this add to the confusion, perhaps leading to protocols in the USA or to national companies selling there new evidenced based solution to the insurance companies …and when that does not work… justification for a spine fusion for axial pain. (screaming now)

    Still are we not missing a key element of the rehabilitation process as Mary says- the multidimensional disorder.
    We do have to get the physical training reasonable right, but don’t we also have to connect with the patient and get “buy in”… and along the way listen to whatever comes up on there minds that are keeping them from caring for themselves, and taking the reins back.

    How do we measure if there is an empathic relationship with a patient and if we have tapped into the other dimensions that are in the room as well?

    EG Reply:

    Hi Mike,

    re: empathy testing,

    But all those tests are a bit ‘sciencey’ for me. When empathy is happening, it’s very obvious to both the therapist and client. You’d never miss it, once experienced.


    Mike Caruso Reply:

    Thanks EG..and so true.

  12. Hmmm… not one subject out of 600 dropped out of the trial? Over 10 years??? Might I suggest a shovel of salt, rather than a pinch, when reading this study?

    Adam Rufa Reply:

    It is also a bit amazing that there is only one author. Usually undertakings like this require significant man power and varying expertise.

    John Ware, PT Reply:

    He didn’t even acknowledge any help in collecting data or writing the manuscript over a ten year data collection period. At the very least, he’s not the most gracious clinical researcher ever to have published an article in Spine.

    Michael Ward Reply:

    Maybe no other author wanted acknowledgment ?

  13. …agree with John unfortunately…

  14. I smell a rat- a disingenuous or at least grossly negligent one. You BiMers are too diplomatic when it comes to reporting on trials like this Aleksiev et al one. And this was published in Spine? How in the world did that happen. Maybe actual rodents performed the peer review?