Low back pain in neurosurgical outpatients: An audit

Low back pain is a common condition. This is a retrospective study of new referrals to neurosurgical outpatients at the Western Hospital, Victoria, Australia. Two hundred and fifteen histories were systematically reviewed. Statistical analysis was performed using univariate and multivariate analyses. Patients who had tried physiotherapy, epidural injection or had no pre-outpatient imaging were more likely to get MRI ( p < 0.02). Patients with clinical features of neurogenic claudication ( p < 0.01) or with neurological signs ( p = 0.02) were more likely to proceed to surgery. CT scan demonstrated significant correlation to MRI for lumbar canal stenosis, disc disease or the absence of disease ( p < 0.01). Referral guidelines for general practitioners regarding back pain are proposed. Recommendations are also made to facilitate the selected use of CT scan and MRI. Keywords Low back pain Referral and consultation Computed tomography Magnetic resonance imaging 1 Introduction Low back pain (LBP) is a term used to describe lumbar spinal pain, sacral spinal pain or lumbosacral pain. 1 It is a common condition, with an estimated lifetime prevalence of 60–80%. 2 While most episodes are self-limiting, the tendency for LBP to undertake a prolonged or relapsing course translates to it being the most common cause of disability across all age groups. 3 This disease burden presents as a significant health problem, and has ramifications for health policies, economic planning and research. The cause of LBP is non-specific in about 95% of people. Serious conditions are rare, with tumours, fractures and infections accounting for less than 1% in general practice settings. 4 Guidelines on the management of both acute and chronic back pain place an emphasis on the safe management of LBP by a careful history and examination without resorting to special investigations. This is a retrospective review of all new referrals to the neurosurgical outpatients at the Western Hospital, Victoria, Australia, for low back pain. Patient parameters are recorded, with a view to identifying clinical variables that influence the request for MRI and surgical intervention. The secondary aim is in correlation of the reporting of disc disease, canal stenosis and nerve root impingement between CT scan and MRI. Referral guidelines for general practitioners regarding low back pain are proposed, aiming to facilitate more appropriate specialist referrals to outpatient clinics. Suggestions for the selected use of CT scan and MRI in LBP are also recommended. 2 Patients and methods 2.1 Inclusion and exclusion criteria The study included all new referrals of patients with LBP presenting to the neurosurgical outpatients at the Western Hospital, Victoria, Australia, between January 2004 and December 2005. Of the 592 new referrals to the clinic, 262 were for LBP. Exclusion criteria were: (i) patients whom after being seen in the public outpatient system, were followed up privately; (ii) patients who were first consulted privately and then referred for follow-up in the public system; and (iii) patients who failed to attend the outpatient appointments, either in the initial visit or in subsequent visits and were thus lost to follow-up. The remaining 215 case histories were then systematically reviewed by one person (JGH). 2.2 Data collection A Microsoft Excel database was set up for data collection. Parameters collected were: patient demographics, waiting time to first outpatient visit, duration of current symptoms, presence of bilateral symptoms, neurogenic claudication, buttock/thigh pain, lower limb symptoms extending below the knee, neurological signs, treatment and imaging modality prior to outpatient visits, and outcome. The data was collected from handwritten notes and typed correspondence. An entry of “sciatica” in the history was categorised as lower limb symptoms extending below the knee. Treatment prior to outpatient clinic attendance was obtained from the referral letter or outpatient notes. Patient outcome was assigned into the following categories: neurosurgical outpatient follow-up, general practitioner follow-up, surgery, medical referrals and allied health referrals. 2.3 Statistical analysis The data was expressed in terms of mean and median using the Microsoft Excel data analysis package. Univariate and multivariate analyses were performed using Stata 7.0 (Stata Corp., College Station, TX, USA). P -values <0.05 were accepted as statistically significant. 3 Results 3.1 Patient characteristics Of the 262 patients referred with LBP to the neurosurgical outpatients clinic, 47 were excluded: 41 failed to attend their appointments and six had their initial or subsequent follow-up privately. The hospital serves the north-western suburbs of Melbourne, where patients were primarily from immigrant backgrounds or low-middle socio-economic class. Table 1 provides a summary of the patient demographics and outcomes. Of the 215 patients in the study, the mean age was 56 years; the median age was 57. The skewness of the distribution was −0.23, in keeping with the general observation that LBP is more prevalent with age. The distribution of gender was in favour of females: 120 females to 95 males. The mean waiting time until first appointment was 12.1 weeks; the median time was 12 weeks. Duration of back pain was classified into two categories: (i) acute 0–12 weeks; and (ii) chronic >12 weeks. 5 In terms of the presentation of pain, 22 patients had LBP only, 37 had LBP with symptoms radiating to the buttock/thigh and 156 had LBP with symptoms extending below the knee. Most patients had tried analgesia prior to their referral, but only 6% of patients had been to physiotherapy. There was no record of other forms of therapy such as hydrotherapy prior to the first outpatient attendance. Following neurosurgical consultations, 78% of patients were managed non-operatively; 22% proceeded to surgery. 13% were referred to the pain clinic for consideration of nerve blocks. 27% were referred for a trial of physiotherapy directly from the clinic; the proportion of patients referred back to general practitioners with recommendations for physiotherapy was not recorded. Less than 1% of patients were referred for hydrotherapy or dietician assessment from outpatients. 3.2 MRI In total, 193 of 215 patients had MRI ordered as part of their work-up for low back pain. Of these, 153 MRIs were ordered after the first consultations in the neurosurgical outpatients clinic; 40 MRIs were ordered pre-outpatient attendance by the referring emergency physicians, other specialists or general practitioners. The patients were divided into those who had a MRI ordered through outpatients (153 cases) and those who had no MRI ordered through outpatients (22 cases). Multivariate analysis was performed. Table 2 shows the likelihood of patients having an MRI if the variables were present. Patients who had tried physiotherapy ( p < 0.01), epidural injection ( p < 0.01), or who had no imaging prior to outpatients ( p = 0.02), were more likely to have a MRI. There were 147 patients who had a CT scan as part of their work-up, and who went on to have an MRI after their initial neurosurgical consultation. The correlation between CT scan and MRI reporting was analysed using chi-square, as shown in Table 3 . There was good consistency between CT scan and MRI on disc disease, canal stenosis or the absence of disease ( p < 0.01). 3.3 Surgery In subsequent analysis, the patients were divided into those who proceeded to surgery and those who were managed non-operatively. Multivariate analysis was performed on the two groups, and is presented in Table 4 . Patients with neurogenic claudication ( p < 0.01) or neurological signs ( p = 0.02) were more likely to proceed to surgery. A subgroup analysis was performed looking at the 154 patients who had a CT scan performed. Chi-square analysis was performed looking at the relationship between patients who had clinical features of canal stenosis or nerve root impingement confirmed on CT scan, and whether such patients proceeded to MRI or surgery ( Table 5 ). Patients, with clinical features and confirmatory findings on CT scan, were more likely to proceed to surgery ( p < 0.01). There was no such relationship when compared to MRI ( p = 0.27). Here, MRI serves more as a pre-operative planning tool, rather than as a diagnostic tool. 4 Discussion Patients included in the study encompassed all new referrals seen in a specialist outpatient setting between January 2004 and December 2005. This group of patients represents a proportion of the patients with LBP in the community. Patient demographic analysis demonstrated similar characteristics to risk factors for the development of LBP: female and age. Other predictors, not measured in the study, include lower socio-economic class, lack of physical activity, obesity and smoking. 6 4.1 Guidelines for referrals Despite the massive literature on the epidemiology and natural history of LBP, the exact causes of back pain in most cases remain ill-understood. Eighty percent of patients with acute LBP can expect resumption to normal activities in 6 weeks; by 12 weeks, this proportion rises to 90%. 7,8 The major issue that confounds clinicians is whether the patient will respond to conservative care or whether surgery is appropriate after failure of conservative management. 9 Management of LBP should involve analgesia and physiotherapy during the initial 12 weeks, both of which can be initiated by general practitioners. Only when a conservative approach fails, should referral to an appropriate outpatient clinic be initiated, so that surgeons can assess if back surgery is indicated. In patients with LBP alone, without any lower limb symptoms, conservative management should persist and referrals for surgical assessment is inappropriate. In our study, 22 patients were referred with LBP alone; all were managed non-operatively ( p < 0.01). Fifty-seven of the 215 patients (27%) in our study had either LBP only or were referred within 12 weeks of symptom onset. Following these simple recommendations ( Table 6 ) would relieve much of the burden placed on the public outpatient system. Waiting time in neurosurgical outpatient clinics would reduce, and there would be a shift toward more appropriate patient selection to the clinics. 4.2 Indications for imaging The routine use of X-rays in LBP makes no difference to outcomes for back pain and has a limited accuracy in excluding serious causes of back pain. 10 The general practitioner therefore needs to consider the rationale behind exposing the patient to the radiation of X-rays for minimal yield, rather than utilise X-rays to achieve false patient reassurance. Many patients referred to the neurosurgical outpatient clinic present with a referral letter requesting management of a patient simply detailed as having back pain and a scan showing canal stenosis. Such brief referrals often contain no information to suggest that a clinical assessment has been performed, which if so, is an inappropriate level of care. CT scan may be ordered for patients suspected of having lumbar canal stenosis with neurogenic claudication after full clinical history and examination, as the study demonstrates significant consistency between CT scan and MRI findings ( Table 7 ). Unnecessary radiological investigations expose patients to risk without benefit and are a threat to the effective allocation of resources. 11 Patients with radiological evidence of canal stenosis, who have LBP alone without features to suggest neurogenic claudication, do not proceed to surgery; the radiology findings here are incidental findings that are not reflected in the clinical condition ( Fig. 1 ). Patients with symptoms and signs of nerve root impingement should be referred onto neurosurgeons directly after failed conservative treatment, so that an MRI can be obtained as the initial imaging modality. CT scan and MRI agreement was statistically not significant in the evaluation of nerve root entrapment ( p = 0.07), and the use of CT scanning exposes patients to unnecessary radiation doses if the ultimate goal of imaging is in pre-operative planning. In 21 patients with CT scans showing no disc disease, canal stenosis and nerve root impingement, MRI has been shown to statistically demonstrate similar negative findings ( p < 0.01). These patients do not benefit from referrals to neurosurgeons or MRI, and should persist with conservative management. None of the 21 patients proceeded to surgery ( p < 0.01). 4.3 Indications for MRI There is little agreement within and among specialities as to the use or prognostic value of imaging findings. 12 The challenge for clinicians, both general practitioners and neurosurgeons alike, is to increase satisfaction without resorting unnecessarily to radiological imaging. 13 Early MRI does not improve clinical outcome for LBP. 14 Much of the anatomical details revealed in an MRI are incidental findings, irrelevant to subsequent clinical management. MRI can serve as the primary imaging modality for diagnosing nerve root impingement from disc, synovial cyst or other pathologies. For patients with clinical features suggestive of central canal stenosis or disc disease, CT scan may still be used as the initial diagnostic radiological tool. For patients who have failed conservative management, and are being assessed for their suitability for surgery, MRI has an important role in pre-operative planning. 9 MRI has no value in planning conservative care 9 and should not be used in the sub-group of patients who are medically unfit for surgery. Neither baseline MRI findings nor MRI changes over time are useful predictors of the subsequent development or resolution of LBP. 15 The indications for MRI are summarised in Table 8 . 4.4 Surgery Forty-seven of 215 new patients attending the neurosurgical outpatient clinic proceed to back surgery. As shown in Table 9 , over half of the patients had neurogenic claudication with central canal stenosis as the underlying pathology, warranting decompressive surgery ( Fig. 2 ). Thirteen patients had disc prolapse, 10 had nerve root canal stenosis and two had thoracic level tumours. 4.5 Observed clinical practice This study provides an overview of patient management in the neurosurgical outpatient setting. The difficulty in distinguishing somatic referred pain from radicular pain is the chief reason clinicians order MRIs, appropriately or otherwise, for LBP. 4 Apart from neurogenic claudication, clinical variables used to distinguish somatic referred pain from radicular pain ( Table 10 ) have been shown to have no significant correlation with the neurosurgeon’s request for MRI. The same pattern has been observed when the clinical variables are correlated with surgery. Of the 215 patients seen, two patients had spinal tumour (0.9%), both in the thoracic region. Both patients presented with atypical LBP and it was the suspicious history and examination findings that led to the incorporation of thoracic imaging. This audit examines the neurosurgical management of LBP from a single teaching institution in Melbourne, Victoria. Its catchment area incorporates the north-western districts of Melbourne, where the patient demographics consist of people from immigrant backgrounds or low-middle socioeconomic class. Generalisation of study findings to other neurosurgical centres in Australia is thus limited. The study examined only patients referred to the neurosurgical outpatients clinic. It acknowledges that the true burden of LBP in the community setting is unclear; the proportion of the patients managed by general practitioners with a normal CT scan and therefore not referred on for specialist opinion is unknown. Out of the 91% of patients who had an MRI in the work-up of their LBP and consultation with a neurosurgeon, 22% proceeded to surgery. Given that MRI has a limited role in conservative management of LBP, further comparative studies could look into whether clinical guidelines on MRI rationalisation may bring about a reduction in the use of MRI in specialist settings. 5 Conclusion In conclusion, LBP is a common condition that presents as a significant health burden to both individuals and society alike. General practitioners play a critical role in the conservative treatment of LBP. Selective careful referrals of patients for surgical assessment after the failure of conservative measures, and the more judicious use of radiological imaging, would bring about a relief to public health resources. This would serve to improve access for patients needing surgical intervention. References [1] H. Merskey The taxonomy of pain Med Clin North Am 91 2007 13 20 [2] N. Bogduk B. McGuirk Medical Management of Acute and Chronic Low Back Pain. An Evidence Based Approach 2002 Elsevier Amsterdam [3] L.C. Giles I.D. Cameron M. Crotty Disability in older Australians: projections for 2006–2031 Med J Aust 179 2003 130 133 [4] V. Wilk Acute low back pain: assessment and management Aust Fam Physician 33 2004 403 407 [5] H. Merskey N. Bogduk Classification of Chronic Pain. Descriptions of Chronic Pain Syndrome and Definitions of Pain Terms 2nd ed. 1994 IASP Press Seattle [6] G.J. Macfarlane G.T. Jones P.C. Hannaford Managing low back pain presenting to primary care: where do we go from here? Pain 122 2006 219 222 [7] N. Boos N. Semmer A. Elfering Natural history of individuals with asymptomatic disc abnormalities in magnetic resonance imaging: predictors of low back pain-related medical consultation and work incapacity Spine 25 2000 1484 1492 [8] L. Abenhaim M. Rossignol D. Gobeille The prognostic consequences in the making of the initial medical diagnosis of work-related back injuries Spine 20 1995 791 795 [9] M.T. Modic N.A. Obuchowski J.S. Ross Acute low back pain and radiculopathy: MR imaging findings and their prognostic role and effect on outcome Radiology 237 2005 597 604 [10] M. Yelland Diagnostic imaging for back pain Aust Fam Physician 33 2004 415 419 [11] R.M. Mendelson C.P. Murray Towards the appropriate use of diagnostic imaging Med J Aust 187 2007 5 6 [12] D.C. Cherkin R.A. Deyo K. Wheeler Physician variation in diagnostic testing for low back pain. Who you see is what you get Arthritis Rheum 37 1994 15 22 [13] D. Kendrick K. Fielding E. Bentley Radiography of the lumbar spine in primary care patients with low back pain: randomised controlled trial BMJ 322 2001 400 405 [14] E.J. Carragee Clinical practice. Persistent low back pain N Engl J Med 352 2005 1891 1898 [15] E.J. Carragee T.F. Alamin J.L. Miller Discographic, MRI and psychosocial determinants of low back pain disability and remission: a prospective study in subjects with benign persistent back pain Spine J 5 2005 24 35