Estimated reading time: 8 minutes
The diagnosis of cauda equina in the emergency department is complex and has the potential for significant patient harm if we get it wrong. Linda Dykes put together an excellent infographic on this back in 2019 that highlighted the updated guidelines for early suspicion and early MR scanning to confirm/refute the diagnosis.
I remember thinking at the time that it was aspirational for many as the access to 24/7 MR scanning in the UK is not where I or anyone else would want it to be. Even in the major trauma centre that is Virchester University Hospital we regularly have to wait until normal working hours for a CT scan. It’s a situation that gives as a lot of anxiety as we know that this is a time critical condition where early decompression saves function.
Whilst MR is an important test, we still need to understand what triggers the need to get the MR scan. No doubt many of us will have had conversations with the MR gatekeeper about whether the patient has the right constellation of symptoms/signs to warrant a scan, and also to prioritise the scan. Such conversations can be difficult, and more importantly quite variable. One scenario described to me from anther hospital is the radiology team declining to do a scan unless approved by the spinal surgeons, whilst the spinal surgeons decline to discuss the patient until the results of the MR scan are known. Such situations are deeply frustrating and unpleasant for EM teams.
Clarity about the signs and symptoms of cauda equina would of course be very helpful in guiding the need and urgency of MR scanning and so when new data is published it’s worth a look. This month we have an interesting paper that challenges some of the traditional signs/symptoms of cauda equina syndrome which might help our patients and ourselves. The abstract is below, but as always please read the full paper and come to your own conclusions.
What type of paper is this?
It’s a retrospective case note review, which is not ideal. Retrospective reviews rely on high quality data entry at the time and also a reliable data storage system. In general studies like this should be done prospectively to ensure that data is collected and stored as accurately as possible. This is slightly mitigated by the fact that all referrals to neurosurgery at this time were recorded on a structured electronic referral system.
Who has been studied?
The ‘cases’ in this paper are those that presented to a single emergency department during a 4-year period. The hospital is also the tertiary neuro/spine centre for the region. This may mean that the case mix is different to that in other emergency departments as patients at risk of cauda equina syndrome may have been preferentially directed to this emergency department by prehospital services, general practitioners, or through local knowledge.
Only patients who had an MR scan were included in the study.
All patients that could be identified as having been referred as possible cauda equina were recorded on the database and used for analysis.
How was cauda equina defined?
Interestingly the gold standard diagnosis here was on the basis of the MR scan rather than operative findings. Although this standard is probably the best available here, it is different from the alternative which would be cauda equina requiring operative intervention. In the study these two events were very similar with only 2 patients not proceeding to operation.
How was the data handled?
As with many papers of this type the data was analysed using univariate and then multivariate logistic regression to determine risk factor.
What did they find?
In the study period there were 996 patients who met the inclusion criteria, of whom 111 (11.1%) had cauda equina on MR scan. A positive rate of around 10% seems about right to me, based on our experience locally. This does mean that about 9/10 scans do not show cauda equina syndrome, but that positive rate, for such a serious condition, appears reasonable.
The results of the study are expressed in a number of formats, which can be tricky to follow. For patient reported symptoms then the following were more likely with a diagnosis of CES.
- bilateral leg plan +/- back pain
- bilateral weakness
For objective signs the following were more likely with a diagnosis of CES
- sensory loss in a dermatomal pattern
- absent ankle jerks
- Upgoing plantar (but small numbers)
In the multivartiate analysis three features indicated a higher risk for the presence of radiological Cauda Equina Compression (CEC) as opposed to no CEC
- Bilateral pain OR 1.9 (CI 1.2-3), p=0.006
- Sensory loss in a dermatomal distribution OR 1.1 (CI 1.1-2.7), p=0.01
- Bilateral ankle±knee jerk reflexes (absent) OR 3.4 (CI 1.8-6.6), p<0.0001
When the individual clinical tests are described for sensitivity and specificity then the individual performance of all the tests was poor, with the highest sensitivity being 50% for loss of perineal sensation. Specificity was only in clinically useful levels for loss of ankle jerks. Likelihood ratios were similarly disappointing, particularly with reference to negative likelihood ratios.
The bottom line for me from the results is that we should not rely on individual clinical signs or symptoms to diagnose or exclude the diagnosis of CES. The best indicators are bilateral leg pain (with or without back pain), objective sensory loss (in a dermatomal distribution) and loss of bilateral ankle jerks.
I’m also interested to see the results regarding digital rectal examination. On the evidence presented here it really has very little diagnostic value, with no difference in findings between those with and without CES. As the authors point out this is in keeping with other studies in the area.
Similarly the use of bladder scanning seems to be of little diagnostic value. Whilst volumes tend to be higher in those with CES there are significant numbers of patients with low and high bladder volumes who both do and don’t have CES.
Should we believe the results?
It’s certainly an interesting paper. It questions the utility of some invasive aspects of clinical assessment (such as digital rectal examination), which is arguably a good thing. A DRE is uncomfortable and intrusive. If we are going to perform it routinely in cases of atraumatic back pain, we need to be clear about how it will affect our management. However, there are lots of limitations. The retrospective and single centre design is always going to be a concern, as is the selection bias of only including patients who had an MR at a neuro/spinal centre. In addition, even with a structured proforma, absent data fields in the project have led to unequal numbers – not every patient appeared to have a documented complete neurological assessment, therefore diagnostic test characteristics for different aspects of assessment are not based on the same total population. All these issues may introduce bias into the results.
That said, there is information here that is of value. Perhaps the most valuable is that the absence of perineal dysfunction cannot rule out cauda equina syndrome. This finding also ties in with recent description of the difference between white and red flags in this challenging condition. Previous authors have reported shifting focus away from ‘white flags’ of surrender and permanent irreversible damage in CES, to more of a focus on genuine ‘red flags’ of warning that may allow timely diagnosis and intervention prior to established neurological injury. An evidence based shift in practice in this area may avoid early MR scans being declined,owing to the absence of specific hard neurological signs.
It’s really important to realise that early MR in suspected cases is almost certainly the key to better patient outcomes. Once patients have irreversible signs (as a result of neurological damage) then we are simply too late. At the moment many centres in the UK cannot deliver timely MR scans for suspected cases and that has to change.
My principal concern with acting on the results of this study is that it is single centre, retrospective and unvalidated. Considering the importance of making this diagnosis in practice we need more research in this area. Of note, the authors entirely agree with this – the lead author Michelle Angus has recently secured funding through a predoctoral NIHR bridging application and intends to use this time to work up a prospective observational cohort study. This is fantastic to hear and we look forward to those findings.
The bottom line.
Patients with atraumatic back pain who are clinically suspected to have cauda equina syndrome following thorough assessment,should undergo MR imaging to evaluate for evidence of radiological compression. Clinicians should not solely rely on the presence or absence of specific clinical symptoms or signs to rule in or rule out CES.
A key message is that the absence of perineal signs should not be used as a reason to not perform an MR scan.
In addition, these findings suggest EM clinicians should carefully consider how a digital rectal examination on patients with suspected CES will influence clinical decision making. This information should also be discussed with patients, to allow informed consent. In the cohort presented, this invasive aspect of routine assessment appeared to add no additional diagnostic value. If these findings can be replicated and validated through prospective research, routine invasive DRE could potentially be avoided without impact on diagnosis.
Simon Carley @EMManchester
with additional insights and contributions from Dan Horner @RCEMProf
Determination of potential risk characteristics for cauda equina compression in emergency department
patients presenting with atraumatic back pain: a 4- year retrospective cohort analysis within a tertiary
referral neurosciences centre https://emj.bmj.com/content/early/2021/10/11/emermed-2020-210540
Does rectal examination have any value in the clinical diagnosis of cauda equina syndrome? https://pubmed.ncbi.nlm.nih.gov/23113877/
Cauda equina syndrome: what is the correlation between clinical assessment and MRI scanning? https://pubmed.ncbi.nlm.nih.gov/17453789/
The accuracy of clinical symptoms in detecting cauda equina syndrome in patients undergoing acute MRI of the spine. https://pubmed.ncbi.nlm.nih.gov/26306934/
CAN 6 Prospective and Retrospective studies https://www.stemlynspodcast.org/e/can-6-retrospective-and-prospective-studies/
Guidelines for cauda equina syndrome. Red flags and white flags. Systematic review and implications for triage. https://pubmed.ncbi.nlm.nih.gov/28637110/