A joint BSCN/EPTA Statement on Hand Held Devices for Carpal Tunnel Testing

A joint BSCN/EPTA Statement on Hand Held Devices for Carpal Tunnel Testing.

20th August 2007

Summary.

The British Society for Clinical Neurophysiology (BSCN) and Electrophysiology Technologists Association (EPTA), in common with some other expert societies round the world, believe that the use of hand held devices for the testing of carpal tunnel syndrome has the potential for significant harm for patients. Such devices should be used only by those who have demonstrated competence in recording and interpreting clinical neurophysiological data. The societies also have doubts about the advantages of such machines over conventional ones.

a. The size of the problem in CTS diagnostics and treatment.

There are approximately 600,000 new cases of carpal tunnel syndrome (CTS) per year in the UK. Most are managed by orthopaedic consultants who, usually, ask for nerve conduction studies on a subset of cases, in which the diagnosis is uncertain clinically. Some consultants do ask for nerve conduction studies (NCS) in all cases for medico-legal reasons. NCS in CTS has historically been performed within departments of clinical neurophysiology. In the UK NCS for CTS makes up approximately 50% of the workload in peripheral nerve testing and this is now usually performed by CPNs though reported and supervised by consultants. In most areas and where best practice is in place there are no long waits for these test, in all probability most already meet the new 6 week target required for March 08. Though the most frequent test performed it is also one of the most important since the results and interpretation have a large impact on patient management.

BSCN and EPTA have agreed a set of minimum standards for CTS testing to improve standardisation and support the extended role of clinical physiologists. These standards are available on our website and a shortened version is attached as appendix 1. In connection with this we also made recommendations about levels of practice which are attached as appendix 2. These recommendations would be contravened in the case of hand held devices if their data were not recorded and reported by a professional trained in neurophysiology.

b. ‘Hand held’ devices.

There is a long history of machines which are marketed to test for CTS alone in a primary care or industrial setting, where little or no expertise in clinical neurophysiology and nerve testing is assumed. These have repeatedly been found inadequate by learned bodies such as the American Academy of Neuromuscular and Electrodiagnostic Medicine, (see references). Most of these older machines used distal motor latency (DML) as the outcome measure. More recently the major manufacturers of equipment in this area have not sought to develop hand held or simplified devices since, in collaboration with clinical neurophysiologists in a number of countries, the limitations and inflexibility of such machines have been manifest.

One machine is now being marketed as a new cheap method for diagnosing CTS in a non-specialist setting. It uses sensory nerve conduction from the digits alone, making it more sensitive than the other, DML based, machines. In a recent paper, apparently impressive results from use of the machine are given (Tolonen et al, 2007), with 85% of those with CTS from traditional NCS testing being picked up. Limitations, accepted by the authors, included the missing of non CTS conditions.

Shortcomings in this machine include;

It measures peak rather than onset latency of the sensory potential.
Peaks are displayed but may not be modifiable.
In testing sensory conduction alone no motor distal motor latencies are included. This machine therefore does not meet the minimal standards set by the BSCN/EPTA. The purpose of testing both sensory and motor nerves is to allow grading of the severity of the CTS which has important implications for management.
The machine is marketed as having the ‘advantage’ of not requiring training in clinical neurophysiology. No recorded waveforms are available to scrutinise but it is specified that the waveforms should be reported by a ‘physician’.

One of the people with the most experience in the UK of management of CTS and NCS testing is probably Dr Jeremy Bland from Canterbury. His comments are as follows;

‘It will work fairly well if users are well trained and the patient population consists of idiopathic CTS of mild to severe severity. It will fail in patients with coincident ulnar nerve problems or generalised neuropathy, it will not give an adequate assessment of the severity of CTS by my standards because no motor study is performed and it is probably only marginally cheaper overall than NCS performed by a clinical physiologist using conventional methods as the major determinant of cost is staff time, not the machine. As such I think the disadvantages outlined above outweigh the small advantage of portability and usability in the orthopaedic clinic. Of course they compare themselves favourably with conventional nerve conduction studies INCLUDING needle EMG examination.... but needle EMG is unnecessary in normal neurophysiological assessment of CTS.’

Further problems include that even in the most carefully selected patients, those attending CTS clinics have been found to have other diagnoses. In one study ‘CTS patients’ were found to normal in 30%, to be positive for CTS in 53% and for ulnar neuropathy in 5% and to have a neuropathy in 11%. With all these machines there is also the problem on false positives, in which the neurophysiological findings of CTS are shown, but in whom other pathology is the cause of the symptoms. One drawback of the Mediracer is that it does not allow the flexibility of further testing available with more conventional machines. For that reason Dr Bland has commented;

‘All such devices are heavily dependent on the knowledge of the people using them and those with sufficient knowledge to understand their limited clinical utility will probably be frustrated when they are unable to extend the testing to meet the clinical need because the only machine they have available is one of these.

Given that the cost of a real EMG machine for performing nerve conduction studies is now down to about £10k or so I think a more flexible device is a much better buy.’

A more detailed response is attached as Appendix 3. In addition another consultant wrote that,

‘In a small but significant number of carpal tunnel decompressions the motor branch may be damaged; without pre-operative motor data this event would not be identified with the use of the hand-held device. If theses cases came to court (as they can do) the use of the hand-held device would be deemed inappropriate, particularly as it does not meet the minimum standards) set down by the Professional bodies.

This is not simply a UK concern. In the US the machine has provoked similar concern and indeed been the subject of an article in the New York Times and concerned the AANEM (New York Times, October 20, 2006). The article expressed concern that doctors would charge for a test rather than for a clinical consultation in a fee led system. Our concerns here are that this machine will reduce the clinical opinion within the patient pathway in CTS and lead to treatments based on inadequate clinical and electrical investigation.

f. Conclusion.

It is difficult to over estimate the concern about this development within the clinical neurophysiology community. All are committed to improving the service but there is a strong feeling that hand held devices, which fail to meet BSCN/EPTA standards, used outside an appropriate environment, have the potential to change patient management in ways with significant potential clinical consequences.

Professor Jonathan Cole, Ms Evadne Cookman,

President, BSCN Chair, EPTA.

This was prepared in collaboration with senior members of both organizations.

References.

Katz RT. NC-stat as a screening tool for carpal tunnel syndrome in industrial workers. J Occup Environ Med. 2006 Apr;48(4):414-8.

Automated Nerve Conduction Testing from the Blue Cross of California. Medicine 0092 05-07-07, which contains a good review of the literature.

Literature review: Nervepace digital electroneurometer in the diagnosis of carpal tunnel syndrome. Muscle & Nerve, Volume 27, Issue 3, 2003. Pages: 378-385.

Tolonen, U, Kallio, M, Ryhanen J et al. A handheld nerve conduction measuring device in carpal tunnel syndrome. Acta Neurol Scand, 2007, 115, 390-397.

J. Thomas Megerian, Xuan Kong, and Shai N. Gozani. Utility of Nerve Conduction Studies for Carpal Tunnel Syndrome by Family Medicine, Primary Care, and Internal Medicine Physicians. JABFM January–February 2007 Vol. 20 No. 1, 60-64.

http://www.aanem.org/PracticeIssues/legislativeAdvocacy/AdvocacyNewsArchives.cfm

http://www.aanem.org/practiceissues/technologyreviews/technologyreviews.cfm

Appendix 1. Standards and Options for CTS investigation.

The following standards are only appropriate for cases selected for the verification of suspected CTS, not for investigation of a differential diagnosis

Standard 1

Before starting testing the patient is identified and the clinical information from the referral verified

Standard 2

Hand temperature is measured, recorded and maintained above 30oC

Standard 3

Sensory nerve conduction is performed on a median digital sensory nerve in the most affected hand using surface electrodes and measuring response amplitude and latency/velocity. A comparative test of conduction in a digital nerve not innervated by the median nerve is performed in the same hand

Standard 4

A test of median motor nerve conduction across the wrist in the affected hand is performed using surface electrodes and measuring response amplitude and latency/velocity

Standard 5

Median motor nerve conduction in the forearm is performed on the affected limb using surface electrodes and measuring response amplitude and latency/conduction velocity

Standard 6

The report of the investigation contains the numerical data. It makes a statement on any abnormality detected. The professional status of the practitioner performing the investigation and report is identified

Standard 7

The report is signed by the practitioner taking medico-legal responsibility for it

Guideline 1

Referrals are screened before allocation of patients by a suitably qualified practitioner to assess appropriateness of clinical question posed

Guideline 2

A focussed patient history and examination are recorded, including the presence of co-existing disease

Guideline 3

Sensory digital nerve conduction as per standard 4 is performed in the contra-lateral hand

Guideline 4

A second test of median sensory nerve conduction is performed. This may include: Median palmar sensory study; Median/Ulnar palmar ratio; Median/Radial sensory latency comparison to thumb; Median/Ulnar sensory latency comparison to ring finger.

Guideline 5

Motor nerve conduction in the ulnar nerve is performed in the affected limb using surface electrodes and measuring response amplitude and latency/conduction velocity

Guideline 6

Median motor nerve conduction is performed in the contra-lateral limb as in standard 4

Guideline 7

The patient is seen by a suitably qualified practitioner at the end of the test to verify the clinical presentation, make a clinico-electrophysiological correlation, to include this in the final report, and to answer any clinical questions the patient may have

Guideline 8

The report details any technical factor that could influence the results.

Statement of Practice in Clinical Neurophysiology, prepared for the PMLG comprising members of the BSCN and the EPTA.

Version; post PMLG April 16^th, 2006.

The Department of Health’s 18 week initiative offers a great opportunity for practitioners in Clinical Neurophysiology to improve their working practice as they seek ways to reduce waiting times. The aims of the PMLG Clinical Neurophysiology have been to conduct a census of waiting times and working practice in England and Wales, to suggest standards and guidelines for investigation and referral which it hopes will become accepted and adopted in England and Wales and to agree clinical competencies and educational requirements for circulation for agreement within the specialty. It is hoped that, once these proposals are accepted, the specialty will be more uniform and efficient.

One way waiting times may be reduced will involve extended working by Clinical Physiologists (Neurophysiology) (CPNs). Though wishing to explore this model, the PMLG is concerned that the nature of the specialty, as part of clinical practice, should remain, with clinical assessment of patients, led by physicians trained and accredited in Clinical Neurophysiology, remaining an important part of the process. This requires working relationships between Consultant Clinical Neurophysiologists (CCNs) and CPNs. In relation to these the PMLG has identified four levels, or patterns, of working;

I. Department of Clinical Neurophysiology with CPNs and full-time CCNs, allowing continual synergy between the two.

II. Department of Clinical Neurophysiology with CPNs and attendance by a Consultant Clinical Neurophysiologist(s) for some days per week. This might be on a hub and spoke model working between neurological centre and DGHs.

III. Department of Clinical Neurophysiology without any attendance by Consultant Clinical Neurophysiologist(s) but with remote supervision using IT.

IV. Department without any input by any CCN. Such a department might be of Clinical Neurophysiology with highly specialist CPNs but might also involve members of a Department of Medical Physics or other health care science disciplines, with tests being performed by practitioners unrecognised by either the BSCN or EPTA.

The PMLG recommends level I and II. It recognises level III practice, but cannot recommend level IV, especially without CPNs.

Appendix 3; based on a letter submitted to Acta Neurologica Scandinavica from Drs David Allen, Ramamurthy Arunachalam, and Prof Kerry Mills. Southampton General Hospital and King’s College Hospital, UK. The authors will have 6 weeks to reply if publication is considered.

28/7/07

Dear Editor –in –chief,

Re: A handheld nerve conduction measuring device in carpal tunnel syndrome.

U Tolonen et al., Acta Neurol Scand 2007; 115: 390-397

We read with interest the paper from Tolonen and colleagues (Acta Neurol Scand 2007; 115: 390-397). Carpal tunnel syndrome (CTS) can be suspected clinically to variable degrees and further supported by nerve conduction studies of the sensory and motor function of the median nerve across the carpal tunnel segment. There is however variability in the clinical presentation and other conditions can mimic CTS. The sensitivity of clinical diagnosis and neurophysiological diagnosis is unknown. Nerve conduction studies (NCS) allow the clinical diagnosis to be confirmed (when positive), allow grading of severity and allow for other conditions to be actively excluded.

New methods and technologies are always welcomed to enhance the diagnosis of carpal tunnel syndrome. Simpler methods and equipment however, should always be weighed up against potential risks in terms of misdiagnosis and missed diagnosis. We praise the authors for their efforts and encourage their continued work in this important area.

The authors sought to assess the diagnostic utility of a device that they themselves have developed. The paper is at times confusingly written, making it difficult to follow through what exactly happened to subjects. We have attempted to understand the results as best we can. The fact that the authors funded, performed and wrote the study, by means of the EMG Technologies Ltd., does raise concern; although this fact is openly disclosed in the paper. There are however also methodological and statistical aspects to the paper, which we feel make it flawed.

Firstly, although the paper seeks to compare the new device and method, to traditional NCS in patients with suspected CTS, it is not explicitly stated how the diagnosis of CTS is made, clinically or neurophysiologically. We have presumed that this was clinically suspected and then confirmed by conventional NCS. They have not compared patients with ‘suspected CTS’ as would be encountered in the clinic but have excluded certain patients from the study. They have excluded those with prior CTS surgery but more importantly those with a history or with NCS findings of any other neurological disorder that may produce numbness or paraesthesia in the hand. If we have understood this correctly, the subjects studied are a highly selected group.

This does not make a direct comparison of NCS and the new device invalid per se, but it does invalidate any attempt to extrapolate the findings to a clinical setting, where the case mix would be far different, and where initial NCS screening would not be undertaken.

It would be anticipated that two machines performing the same function, that is sensory nerve conduction studies, in a highly selected patient group, would needless to say, result in a high level of concordance between results. Despite this the new device is neither as specific nor as sensitive as traditional NCS. Its use also raises concerns regarding potential patient misdiagnosis. For example the device, as the authors state, would miss ulnar neuropathies, a condition not infrequently referred to clinics as ‘CTS’. This diagnosis would be picked up by a traditional dynamic, clinically driven NCS/EMG examination, which would respond and alter course according to the findings as they were found. In the case of the new device performed by non-neurophysiologically, non-medically trained personnel, the patient would be classed as not having CTS, which would necessitate possibly two further appointments, one for a further clinical examination and then further NCS studies. This would clearly be uneconomical.

There are several statistical and data analysis points that we feel require clarification. Firstly, the terms sensitivity and specificity are incorrectly used.¹ The authors report (p393) that the ‘sensitivity of the new device is 97.3%’. But the sensitivity, is actually the proportion of patients with CTS who are correctly diagnosed by using the new device. They are referring to the specificity. The sensitivity of the device therefore appears to be 85.5% according to their figures (171/200) and it is less (80.8%) when the automated program (for non specialist users) is used. This amounts to a false negative rate of about 1 in every 5 patients. Again the specificity would be the proportion of true negatives who are correctly identified by the test, and not the proportion of false positives.

On several occasions the figures do not appear to add up when attempting to follow the flow of subjects and results. Of 194 suspected CTS cases 129 had CTS confirmed on NCS, 78 of these had it in both hands. This equals to 207 hands. This figure is then not directly further discussed or broken down. Later on, 200 hands are referred to, leaving 7 unaccounted for. Some of the values in Table 3 also appear to be incorrect. The percentages in parenthesis do not relate to some of the preceding values given e.g. in the mild (R) CTS group 27 out of 36 patients were correctly identified by the device. This amounts to 75% and not 86.1%, i.e. a quarter of the patients in this group of mild CTS were not identified with the new device. Similar errors with the figures are made in the moderate group as well.

In the discussion, the opening statement suggests that the new device agreed with traditional NCS in 91% of the cases. It is not clear how this figure has been statistically reached, and is not supported by the data presented. The authors further state that due to unsatisfactory data and missing responses 8% would need to be re-referred to have traditional NCS. In fact, the number that ‘should’ be re-referred for NCS in our opinion is higher. With the new device 44 hands (39 with useful data from new device) had ‘severe’ CTS. Of these none had a bifid response from the ring finger and only 6 had any response from forefinger stimulation. So at least 33 hands in this group, but possibly more, would have been identified as having an absent median nerve sensory response. The new device tests purely sensory function. A diagnosis of CTS cannot be confirmed, let alone graded as severe CTS in this instance, without additional motor studies. Potential missed diagnosis would include high median neuropathies and brachial plexus lesions, for example. These patients would also require referral for traditional NCS to safely establish the diagnosis. In total sixteen subjects were excluded, 5 of these from the severe group. So an additional 11 from other groups were excluded due to poor data. It is clear that the number of subjects subsequently requiring referral for traditional NCS is fast approaching about a quarter of the study group. It should be remembered that according to its reported sensitivity, about 15-20% of the group with CTS would also not have been detected with the new device. The device in its current form also appears to perform sub-optimally; digital data was lost during transfer to the main server in 15 patients.

The above refer to patients from a highly selected group. To extrapolate the findings from this study to clinical use would be misleading. A number of conditions can present as suspected CTS. In practice, those referred as ‘suspected CTS’ frequently have features in the history and/or examination that make the diagnosis clinically far from certain, prompting their referral.

The new device also yielded three abnormal responses, where traditional NCS did not. These could be termed false positives, though the authors felt that the patients actually had CTS on clinical grounds. The explicit definition of CTS for the study’s purposes, therefore seems to have altered or have made exceptions during its course. The authors explained that the NCS studies performed included the same studies as those performed by the new device, so why has this difference arisen if they are not false positives.

The authors refer to the neurophysiological grading system devised by Padua et al.²

Here ‘severe = absent median SNAP or profoundly decreased amplitude in addition to a severe prolonged median motor distal motor latency’. Given that only sensory responses are recorded with the new device the authors have had to rely on traditional NCS for grading purposes and have provided mean values for the latencies of sensory potentials in CTS of differing severity. In the severe grade it should be noted that the majority of subjects (38/44) had absent sensory responses

Sufficient comparison of pre and post-operative NCS in CTS, especially in the context of surgical failures, require motor nerve conduction studies. It is therefore logical that in patients for whom a surgical treatment is being contemplated, pre-operative motor NCS should be performed.

We feel that our arguments, although strongly presented, can be justified by considering the potential effects of the use of such a device, when it is used for assessment of a suspected condition with such a high incidence and prevalence. Thousands of patients are seen with suspected CTS. The economic inefficiency of many patients requiring further ‘repeated’ studies with traditional NCS and the questionable safety of such a device in the accurate diagnosis of CTS, should make readers think twice, before accepting such a device.

The real test for such a device would be an independently performed prospective blinded study using an unselected patient group.

CTS should probably be considered as an electro-clinical diagnosis. The NCS studies and the patients require a clinical opinion. Mild or sometimes more severe CTS is all too often picked up incidentally during NCS. It does not equate that this is responsible for the patient’s symptoms. The findings should fit with the symptoms and their severity and other diagnosis considered and excluded where necessary. Patients deserve an appropriate electro-clinical assessment and not just a test in isolation.

Dr David Allen*

Dr Ramamurthy Arunachalam*

Prof Kerry Mills^†

*Southampton University Hospitals NHS Trust; ^†King’s College Hospital, London, U.K.

References:

1. Altman DG, Bland JM. Diagnostic tests 1: sensitivity and specificity. BMJ 1994; 11;308(6943):1552.

2. Padua L, LoMonaco M, Gregori B, Valente EM, Padua R, Tonali P. Neurophysiological classification and sensitivity in 500 carpal tunnel syndrome hands. Acta Neurol Scand 1997; 96:211-7.