What evidence do we need to develop clinical guidelines for spinal fusion in patients with chronic low back pain?


By Paul Willems

Recently, we performed a nationwide survey among spinal surgeons in the Netherlands regarding clinical decision making for spinal fusion in patients with chronic low back pain.1 The survey revealed a complete lack of consensus in the use and appreciation of commonly used tests for patient selection, such as trial immobilisation, orthosis or provocative discography. Prognostic patient factors, such as smoking, obesity, work status or psychosocial problems, were not consistently incorporated in treatment strategies. The considerable heterogeneity in practices could not be explained by differences in training or clinical experience, but were more likely to be because of the absence of a generally acknowledged best of care for low back pain.


In a systematic literature review, we determined the accuracy of the above mentioned tests for patient selection.2 With likelihood ratios approaching one, all tests failed to accurately predict the outcome of spinal fusion. In particular, specificity was low—meaning that for all tests, high proportions of false-positive test results lead to unnecessary invasive and expensive surgery, as those patients with a bad outcome after fusion could not be identified. It was disappointing that for MRI, the recommended evaluation tool for low back pain, not one study could be found that determined its prognostic accuracy for fusion.


The lack of proven accuracy of prognostic tests is reflected in the high degree of clinical uncertainty regarding fusion for low back pain. Our results confirmed that in many clinical practices, patients are scheduled for fusion on the basis of tests of which the accuracy is insufficient or at best unknown. Best evidence does not support the use of these expensive and invasive tests in standard care.


Chronic low back pain is not a diagnosis but rather a symptom in patients with different stages of impairment and disability. It should be acknowledged that social and psychological factors greatly influence its outcome, and thus patients should be screened for these factors, preferably by psychodiagnostic questionnaires. In current medical practice, a patient seeking care can contact physicians from different medical disciplines. These physicians all have their specific expertise and therapy proposals, which may differ substantially. Moreover, there is not even assurance that physicians within the same medical discipline will offer similar treatment. Hence, patients should be evaluated in a multidisciplinary spinal centre setting where clinicians are aware of all commonly used therapies and willing to refer patients to colleagues, who may provide the most appropriate treatment for that patient.


Cognitive behavioural therapy or intensive exercise programmes have outcome results similar to those of spinal fusion,3–5 but with considerably less complications, morbidity and costs.6 Our findings showed that the currently used prognostic tests cannot improve the results of fusion by better patient selection, which makes it hard to propose spinal fusion as a standard treatment for low back pain.


To verify whether spinal fusion could be effective for a subset of patients with persisting symptoms, future research should focus on studies that include both positively and negatively tested patients in a randomised design between fusion and programmed conservative care. Test protocols should be clearly described and clinical outcome should be defined by a consensus cut-off point of improvement in pain and functional status, a so-called minimal clinically important change.Additionally, MRI and patient risk factors for persistent disabling low back pain,8 should be included. Randomised controlled trials may provide a high level of evidence, but are often restricted by a limited external validity. Therefore, complementary observational data with indeed a lower level of evidence but a higher feasibility of being closer to daily practice should be collected additionally.


The implementation of well-organised national and international spinal registries should enable the spinal community to acquire representative data of high validity, to be used for the instalment of consensus guidelines. We, as clinicians, need such guidelines, not only to counsel our patients, but also for communication with other healthcare providers, insurers and policy makers to clarify the role of spinal fusion among other treatment options in the management of chronic low back pain.




1. Willems, et al. BMJ Open 2011. Epub

2. Willems, et al. PhD-Thesis University of Maastricht 2011, ISBN/EAN: 978-90-8590-051-1

3. Brox, et al. Spine 2003; 28:1913-21

4. Fairbank J, et al. BMJ 2005; 330:1233

5. Guzman J, et al. BMJ 2001; 322:1511-6

6. Deyo et al. JAMA 2010; 303: 1259–65

7. Ostelo, et al. Spine 2008; 33:90–4

8. Chou, et al. JAMA 2010;3 03:1295-302


Paul Willems, Department of Orthopaedics, Research School Caphri, Maastricht University Medical Center, The Netherlands.