Which specific modes of exercise training are most effective for treating low back pain?

  1. Exercise is a recommended treatment for chronic low back pain (CLBP). Current guidelines state that no one type of exercise is superior to another, and all have small to modest effects on pain and disability.
  2. This network meta-analysis found that pilates had the largest effect on pain, that resistance and stabilization/motor control exercise had the largest effect on function, and aerobic exercise had the largest effect on mental health; among individuals with CLBP compared to no intervention and non-exercise interventions.
  3. This paper has significant methodological shortcomings which limits its relevance to clinical practice.


A network meta-analysis published this year in the British Journal of Sports Medicine aimed to determine what is the most effective type of exercise for individuals with chronic low back pain (CLBP), in terms of pain, disability, and mental health. This is an important topic for physiotherapists as exercise is a mainstay treatment for chronic low back pain.

Accordingly, this review sought to determine whether some types of exercise are better than others for improving pain, function, and mental health among individuals with CLBP.


The authors conducted a network meta-analysis of randomized controlled trials. They searched five electronic databases from inception to May 2019 to locate potentially relevant studies. There were no language restrictions. The quality of

the evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria. The inclusion criteria were as follows:

Study design – Randomized controlled trials only

Participants – Adults (≥18 years of age) with chronic non-specific LBP (≥12 weeks)

Instead of focusing on the type of exercise for LBP, we should focus our efforts on getting people to do more of any exercise, and on strategies to help individuals stick with exercise over time.

Intervention – Exercise alone for at least 4 weeks duration. Exercise included resistance, stabilization/motor control, pilates, yoga, McKenzie, flexion (controlled movements
in flexion only), aerobic, water-based, or multimodal exercise (a combination of different types of exercise).

Comparison – These interventions could not involve exercise. There were three categories:

  • No intervention provided
  • Hands-on treatment – manual therapy, chiropractic, passive physiotherapy, osteopathic, massage or acupuncture
  • Hands-off treatment – general practitioner management, education, or psychological interventions

Outcomes – Pain intensity (e.g. numerical rating scale [0 -10]), physical function (e.g. Oswestry Disability Index), and mental health (e.g. 36-Item Short Form Health Survey).


  • The authors included 89 trials.
  • The review found low certainty evidence that pilates, stabilization/motor control, resistance training and aerobic exercise were the most effective exercise interventions compared to other exercises, as well as non- exercise comparisons.Exercise is a mainstay treatment for chronic low back pain.
  • Specifically, the review found that pilates was most effective for pain, stabilisation/motor control for function, and aerobic for mental health.


Unfortunately, this review has significant limitations and has been the topic of great controversy since its publication (1). Major limitations include:

  • Missing a lot of relevant trials – it is estimated that the review missed 261 eligible trials.
  • The estimates provided for the effect of pilates on pain are likely implausible – with authors presenting effect sizes about 3-4 times that normally reported for exercise interventions in CLBP.
  • Poor reporting of methods – the authors made statistical decisions without being transparent on how and why they picked these methods.



The limitations of this paper mean that we cannot have confidence in the results found. The highest quality evidence (2) available on exercise interventions for CLBP states that all types of exercise deliver a small to modest effect on pain and disability, and no one exercise is superior to another. We should be guided by this evidence and place less emphasis on the ‘right’ type of exercise. Preoccupation with the right exercise for CLBP needlessly complicates care and could give patients the impression that some exercises are ‘good’, and others are ‘bad’.

Instead of focussing on the type of exercise for LBP, we should focus our efforts on getting people to do more of any exercise, and on strategies to help individuals stick with exercise over time. Depending on the individual, this might involve reducing people’s potential fears about exercise for back pain (e.g. through education combined with exercise), getting people more confident with movement (e.g. exposure, graded activity, encouragement), and always considering a person’s unique preferences (e.g. what activity does a person enjoy or least dislike) and goals.

Leave a comment