Intended for healthcare professionals

Practice Change Page

Avoid surgery as first line treatment for non-specific low back pain

BMJ 2014; 349 doi: (Published 16 July 2014) Cite this as: BMJ 2014;349:g4214
  1. Wilco C Peul, professor and chair of neurosurgery12,
  2. Annelien L Bredenoord, associate professor of biomedical ethics3,
  3. Wilco C H Jacobs, health scientist1
  1. 1Department of Neurosurgery, Leiden University Medical Centre, PO Box 9600, 2300 RC Leiden, Netherlands
  2. 2Department of Neurosurgery, Medical Centre Haaglanden, PO Box 432, 2501 CK The Hague, Netherlands
  3. 3Department of Medical Humanities, Julius Centre, University Medical Centre Utrecht, Netherlands
  1. Correspondence to: W C Peul w.c.peul{at}

Key points

  • For non-specific low back pain (LBP), do not offer spinal surgery as first line treatment, as insufficient evidence exists for greater effectiveness of surgical interventions than conservative treatment

  • Offer surgery only for severe persistent non-specific LBP if multidisciplinary (combined physical and psychological) conservative treatment is ineffective

  • Non-specific LBP should not be confused with sciatica or neurogenic claudication, for which surgery may have a role

  • Magnetic resonance imaging should be performed only if red flags exist or when considering surgery for persistent and severe non-specific LBP for which conservative treatment has been ineffective

Low back pain (LBP) ranks as the number one disorder in terms of years lived with disability; estimated global one year incidences range from 22% to 65%.1 Back pain as a symptom, not attributable to spinal instability caused by trauma, infection, progressive deformity, or tumour, and not associated with radicular symptoms, is labelled non-specific LBP (see box). Many of patients with non-specific LBP may have degenerative intervertebral disc and bony joint changes, including disc herniations and spinal stenosis on imaging, but these findings are not more common than in the general population, including people without LBP.2 Guidelines discourage surgery for non-specific back pain3; some recommend surgery only after two years of failed conservative treatments in carefully selected patients.4 However, despite these recommendations, the rate of back surgery is rising, with considerable geographical variation.5 The rates of spinal fusion, including use of surgical implants, for all degenerative disorders of the lumbar spine show an unexplained exponential increase,6 7 with a corresponding rise in serious perioperative complications, including stroke and cardiopulmonary events.7 These rises are unexpected, as the incidence of spinal diseases causing instability of vertebral elements has not risen, and nor has the incidence of surgery for herniated discs and stenosis.5 We propose that surgery should be avoided as first line treatment of non-specific LBP and that it is justified only in selected well informed patients after multidisciplinary conservative care has failed.

Conditions that may cause low back pain

  • Herniated intervertebral disc—This may cause nerve root compression, which in turn may cause radicular leg pain, low back pain (LBP), or both

  • Isthmic spondylolisthesis—Anterior or posterior displacement of a vertebra relative to the vertebra or sacrum below it, due to a bone defect in the posterior elements of that vertebra; this may cause compression of nerve roots, which in turn may cause LBP, radiating pain, or both

  • Degenerative spondylolisthesis—Anterior or posterior displacement of a vertebra in relation to the vertebra or sacrum below, due to degenerative processes in the discs and articular surfaces of the joints connecting vertebrae; this may cause compression of the spinal cord or nerve roots, which in turn may cause LBP, radiating pain, or both

  • Spinal stenosis—Abnormal narrowing (stenosis) of the spinal canal causing compression of the spinal cord or nerve roots with associated neurological deficit including low back or radicular pain, numbness, paraesthesia, and muscle weakness

  • Non-specific LBP—Pain with possible tension, soreness, or stiffness in the lower back region, for which no specific cause can be identified. Several structures in the back, including the joints, discs, and connective tissues, may contribute to symptoms.3 Imaging findings of disc herniation and spinal stenosis may sometimes be incidental and are not more common than in people without LBP2

The evidence for change

For non-specific LBP, studies repeatedly show that surgery does not offer clinically relevant benefits over conservative interventions such as multidisciplinary treatment. For instance, in non-specific LBP with documented degenerative disc disease, a systematic review compared intradiscal electrothermal therapy with sham surgery.8 This included two randomised controlled trials (121 patients), which both showed no significant difference in outcomes, such as pain, disability, and quality of life. In a meta-analysis that included four higher quality randomised controlled trials (767 patients) comparing spinal fusion surgery with conservative interventions, surgery was not superior to intensive rehabilitation in improving back pain, disability, and quality of life.9 A Cochrane review of five randomised controlled trials with 1301 patients evaluated disc replacement for degenerative disc disease.10 It included one trial at low risk of bias (173 patients) that showed a statistically significant (12.3 mm on a visual analogue scale for back pain) but clinical irrelevant (less than 20 mm) difference favouring disc replacement over rehabilitation. Potential harms or surgical side effects were inconsistently reported in studies. The guideline from the National Institute for Health and Care Excellence (NICE) on non-specific low back pain included only one cost effectiveness analysis, which concluded that the chance that surgery as early treatment is cost effective compared with non-operative care is only 20% if decision makers are willing to pay £30 000 (€37 600; $51 100) per quality adjusted life year.3

However, surgery may have more of a role for relief of radicular pain and LBP with specific causes. For back pain with predominant sciatica due to herniated lumbar discs, a systematic review with five (two with a low risk of bias) randomised controlled trials (1135 patients) concluded that early surgery leads to short term benefits in function and leg pain (recovery after 4 v 12 weeks) but with similar long term results.11 Surgery did not offer superior relief of concomitant LBP in these studies, and patients with predominant LBP were excluded. For back pain with predominant neurogenic claudication due to lumbar stenosis, systematic reviews show that, for patients with neurogenic claudication with or without minimal back pain, bony decompression (five randomised controlled trials with 918 patients) and interspinous devices (two randomised controlled trials with 142 patients) may be better than conservative treatment from three months up to at least four years in improving pain, disability, and quality of life.12 13 14 These studies excluded patients with predominantly LBP as their major complaint. For back pain due to isthmic spondylolisthesis (see box), in a systematic review of eight randomised controlled trials (385 patients),15 only one trial (111 patients) compared surgery with conservative treatment, showing improved overall outcome (74% v 43%) at two years after posterolateral fusion versus exercise, but this trial had a high risk of bias. Unfortunately, for those patients with so called “good” anatomical and physiological indications for surgery (for example, spinal trauma, tumour, infection, progressive deformity), back pain as an outcome was not studied.

Barriers to change

Evidence suggests that surgery may have a role in patients with sciatica from herniated lumbar discs or neurogenic claudication due to spinal stenosis. However, patients and physicians may wrongly extrapolate this to non-specific LBP, especially when pain is poorly controlled. The low threshold for performing magnetic resonance imaging for back pain may also result in the diagnosis of structural abnormalities of questionable clinical significance, triggering treatment algorithms and the quest for surgery. Placebo effects of surgery should also not be underestimated. In addition, the medical device industry expends great effort in promoting surgical procedures with implants, and complications are often under-reported in the literature, especially where financial ties exist between authors and sponsors.16

Reluctance to generate systematically collected evidence to support the rise in surgery for non-specific LBP might be explained by “surgical exceptionalism”—the view that the ethical or regulatory status of surgery is justified by its (questionably) unique nature.17 The view that if interventions do no good they will at least do no harm is a misunderstanding, especially given the potential discomfort for patients and possible risks of surgery.18 Evidence based surgical practice is necessary to protect patients against unproved “treatments” and insufficiently tested devices, as well as to make rational decisions in times of scarcity.

How should we change our practice?

In view of the above, spinal surgery should be avoided as first line treatment for non-specific LBP.19 It should instead be considered only as the last line of treatment for persistent and severe LBP if an adequate trial of multidisciplinary non-operative treatment, including a combined physical and psychological treatment programme, has failed and following adequate informed consent. This is in accordance with the NICE guidance, which found insufficient evidence to advocate surgery as first line treatment for non-specific LBP.3 Physicians should advise patients that, for non-specific LBP, evidence is lacking to show that surgery significantly improves pain, disability, and quality of life or is any better than physiotherapy, multidisciplinary non-operative care, or a careful “wait and see” approach.

Where non-operative measures are not readily available, healthcare organisations should ensure that these are a priority for implementation. Surgical societies and the medical community generally need to promote cultural change towards evidence based spinal surgery and create ample room for informed decision making by the patient during the consultation.

Magnetic resonance imaging can lead to misleading anatomical diagnoses and should not be performed for back pain. It should be considered only when referral is indicated for consideration of surgery. If pain due to sciatica or neurogenic claudication is significant or is impairing function or quality of life, imaging and subsequent surgical procedures should be considered. The indications are more urgent if red flags, such as a neurological deficit or a history of cancer, exist (table).4

Red flags prompting further diagnostic investigations in case of first episode of low back pain

View this table:

Any surgery for non-specific LBP should be done only under controlled study conditions. Surgical innovation outside a randomised controlled trial is not in itself unethical, but the patient should be fully informed about the lack of evidence for added value in terms of reduction in back pain and objective evaluation of clinically important outcomes should be ensured.

Sources and selection criteria

We searched the Cochrane Database of Systematic Reviews, Cochrane Database of Abstracts of Reviews of Effectiveness, Cochrane Central, and PubMed. We selected the most recent, high quality systematic reviews comparing surgery with conservative treatments for the treatment of patients with chronic low back disorders


Cite this as: BMJ 2014;349:g4214


  • Change Page aims to alert clinicians to the immediate need for a change in practice to make it consistent with current evidence. We welcome any suggestions for future articles (email us at practice{at}

  • Contributors: WCP developed the concept of the article. WCHJ did the literature search and developed the first substantive draft. ALB discussed the ethical rationale of evidence based surgery and its implications. All three authors contributed to consecutive and final drafts and read and approved the final version. WCP is the guarantor.

  • Competing interests: We have read and understood the BMJ policy on declaration of interests and declare the following interests: WCHJ received unrestricted research support for investigator initiated clinical studies from ZonMw and AO Spine, is coordinator of the multidisciplinary Dutch network for spinal disorders, and has an unpaid advisory role on an outcome measurement committee for AO Spine, for the guideline on instrumented spine surgery, and for the guideline on spinal metastases; WCHJ and WCP are authors of some of the evidence used in the manuscript; WCP received unrestricted research support for investigator initiated clinical studies from ZonMw, CVZ, Medtronic, Braun Aesculaep, Paradigm Spine, The Hague Hoelen Stichting, and the Leiden University Medical Center. ALB received grants from ZonMw and BMM.

  • Provenance and peer review: Commissioned; externally peer reviewed.


View Abstract