Double Crush Syndrome

The original version of this article was written as a reponse to a question on a bulletin board by a patient and it has been revised for this website. Another very sensible review of the topic by an academic chiropractor can be read online here.

The term ‘double crush syndrome’ originates from work published by Upton and McComas (1973) and as such it would follow established practice if their original statement of the hypothesis were to become the working definition of 'double crush'. The abstract of their paper read as follows:

"A comprehensive electromyographic study has been made of 115 patients with carpal-tunnel syndromes or lesions of the ulnar nerve at the elbow. In 81 cases there was electrophysiological evidence, often supported by clinical symptoms, of associated neural lesions in the neck. This association is not thought to be fortuitous, but rather the result of serial constraints of axoplasmic flow in nerve fibres."

However, I have since had discussions with Dr. Upton in which he has extended the overall concept to include situations which cannot result from ‘serial constraints of axoplasmic flow in nerve fibres’ because he would now like to include within the syndrome cases in which the two lesions actually affect different nerve fibres. Nevertheless the original hypothesis may be stated as:-

Local damage to a nerve at one site along its course may sufficiently impair the overall functioning of the nerve cells that they become more susceptible than would normally be the case to trauma at other sites. It is worth noting however, that for sensory nerve cells, this implies that the proximal component of the double crush needs to be between the dorsal root ganglion (which lies outside the spine) and the carpal tunnel, not between the spinal cord and the dorsal root ganglion, which is where most entrapment from degenerative disease of the cervical spine occurs.

Upton and McComas further suggested that a high proportion (75%) of patients with one peripheral nerve lesion did in fact have a second lesion elsewhere and they implied that both lesions were contributing to the symptoms. 

The idea that patients may have more than one lesion, and even the concept that both may be contributing to the symptoms is, and was not in 1973, anything new. The new elements of Upton and McComas's paper were the idea that one lesion could PREDISPOSE to the other, and the very high percentage of patients in whom they found evidence of two lesions. The element of predisposition by a mechanical lesion remains unproven 38 years later, and in the absence of definite proof on this, the term 'double crush' has widened somewhat to include symptoms which result from a combination of two separate, local lesions at different anatomical sites in the same nerve, whether or not one actually contributes to the causation of the other.

I feel that attempts to widen the definition still further (see my comment about recent discussions with Dr Upton) should be resisted unless there is VERY good reason. Provided we keep more or less to the original proposition the essential elements are then:-

1) This is a neurological problem (i.e. involving nerve only and not a combination of a nerve lesion with pathology in another body system – joint, tendon etc)
2) There are two physically separate lesions of the same nerve
3) Symptoms are present

Anyone using the term 'double crush syndrome' for cases which do not meet these three criteria is misusing the term. It is not strictly a diagnosis but rather, all cases of double crush syndrome should be describable in terms of two other diagnoses.

Is there no more recent research?

Yes, there is plenty, much of it contradictory. When I wrote the original version of this article I quoted 7 papers at this point. My collection now includes 31 papers with something to say about double crush syndrome and there are doubtless more. There have been some attempts to test the hypothesis in animal experiments (Nemoto 1987, Dellon 1991, Seiler 1983,Shimpo 1987) but in general these have only been able to show that two lesions, unsurprisingly, produce a greater overall deficit in nerve function than one. More studies have been devoted to trying to assess the incidence and clinical relevance of the syndrome. In their original paper Upton and McComas claimed that 75% of their patients with CTS or ulnar nerve lesions had evidence of a second proximal lesion. Other studies have come up with the following:

• 1000 hands in 888 patients – 11 felt to have double crush (1%) (Carroll 1982)
• 100 clinically diagnosed CTS patients – 6% thought to have double crush (Kuntzer 1994)
• Double crush in 0.7% of 12736 patients (Morgan 1998)
• 40 of 271 patients with thoracic outlet syndrome felt to have CTS too (Narakas 1990)
• 37 of 1183 patients with CTS found to have cervical radiculopathy (3.1%) (Wilbourn 1986)
• 74 of 165 cases of thoracic outlet syndrome found to have a peripheral entrapment neuropathy (44%) (Wood 1990)
• Evidence of double crush found in 74 limbs out of 758 in 681 patients with thoracic outlet syndrome (9.76%) (Abe 1999)
• Of 525 patients with CTS 105 (20%) had other entrapments in the same limb (Yu 1979)

In general the incidence of double crush in patients with definite CTS has tended to be lower than that suggested by Upton and McComas though there have been claims that it exists in as many as 83% of CTS cases and it is often invoked as a reason for failed carpal tunnel surgery. It does seem however that the rates of CTS in patients with thoracic outlet syndrome may be higher than those of thoracic outlet syndrome in patients with CTS.

How can supposedly scientific data vary so much?

Much of the evidence in these papers is derived from EMG (Electromyography) and nerve conduction studies. To understand the variations one has to understand a good deal about the nature of the tests and the way in which tests in general are used in medicine. One also has to appreciate something of the relationship between disease and symptom prevalence in the population.

EMG and nerve conduction studies are not 'all or nothing' tests producing a logical YES/NO answer as to whether a problem is present or not. With nerve conduction studies, at least the results are numerical measurements and different investigators can agree about what the limits of normal are, but even here there is some overlap between patient groups - some patients with no symptoms will have results that fall in the 'abnormal' range (false positives) and some patients with disease which is causing symptoms will nevertheless have results which are normal (false negatives). There is therefore no absolute dividing line in the test results which will reliably place all patients correctly into the two groups 'Disease' and 'Normal' and one has to set an arbitrary cut-off point as 'normal'. This can be set so as to minimise the number of false positive results or to minimise the number of false negatives. If the consequences of a false negative are catastrophic - for example in other circumstances, missing a potentially treatable malignancy - one tends to bias the cut-off point towards minimising the false negatives at the expense of having rather more false positives and vice versa. One can therefore see that there is no absolute 'RIGHT' answer.

With EMG, the situation is further complicated by the fact that the 'result' is largely a subjective judgement by the examiner rather than a quantifiable number. This is far more difficult to standardise between laboratories and is open to wide interpretative variations, further compounding the difficulties outlined above for nerve conduction studies.

Now what about the relationship of disease to symptoms? The concept of asymptomatic disease is easy enough to comprehend and everyone is now familiar with the idea of screening tests to detect such. Not so familiar however is the fact that some tests, and in this I would include post-mortem studies of accident victims and the like, produce positive results in surprisingly high proportions of the normal population, for example up to 40% of people may show pathological evidence of carpal tunnel syndrome at post-mortem. The physician can thus frequently be presented with the difficult problem of determining whether the symptoms of which his/her patient is complaining are actually due to the disease which he has been able to demonstrate by examination or investigation. In doing this, he or she is forced to fall back on knowledge of what symptom pattern is typically associated with each disease and has to make a judgement as to how closely the patients symptoms match - and by implication, how likely it therefore is that the disease in question is responsible.

How does this apply to RSI, Carpal Tunnel Syndrome (CTS) and Double Crush? Well, two types of pathology are known to be very common in the population at large and therefore present the problems outlined in the last paragraph - these are CTS and the group of conditions which, for want of a better term, I will call 'root/plexus problems'. This last includes arthritic degeneration of the cervical spine producing nerve root compression, thoracic outlet syndromes (cervical ribs) and perhaps entities such as adverse mechanical tension. Now, EMG and/or nerve conduction studies may be used to look for evidence of both of these. The mainstay of examination for CTS is nerve conduction studies, whereas examination for root and plexus problems relies much more heavily on EMG with its attendant difficulties. The end result is that neurophysiological tests for CTS produce a much more reliable (though not perfect) differentiation of 'DISEASED' from 'NORMAL' than neurophysiological tests for root and plexus problems where there is a much greater overlap between the test results of the symptomatic and asymptomatic populations. Given the situation that I have described above, it is not difficult to see that one investigator can inadvertently bias his/her findings towards finding evidence of second lesions in the neck/shoulder and another investigator can do the opposite.

Does 'double crush' explain much misdiagnosis of CTS?

It is sometimes suggested that the failure of symptoms to respond to treatment for CTS, often surgery, is a result of misdiagnosis of double crush syndrome as CTS. The implication is often that many patients who have double crush syndrome are mistakenly treated as having CTS when they don't have it – as though the two conditions are entirely separate. Most of the patients we are talking about will have CTS as one part of their 'double crush' as CTS is far more common than other entrapment neuropathies in the arm. For practical purposes, a double crush consisting of CTS combined with a root/plexus lesion probably accounts for 95% of all double crush syndromes. It may help at this point to enumerate the possible scenarios for an ‘RSI’ patient:

1) The patient has CTS and a plexus/root lesion both contributing symptoms independently
2) The patient has CTS causing symptoms and an asymptomatic root/plexus lesion which is completely unconnected
3) The patient has symptomatic root/plexus disease and an irrelevant and asymptomatic CTS
4) The patient has CTS which would normally cause few or no symptoms alone but the presence of a root/plexus lesion is somehow enhancing the symptoms and making the CTS a problem. In another version of this scenario the patient has raised pressure in the carpal tunnel which would not normally be enough to damage an otherwise unimpaired nerve but because the median nerve is in trouble proximally it is unable to cope with the situation at the wrist and develops a lesion there too. In both versions the symptoms are only those of the CTS otherwise this would be case 1)
5) The patient has a root/plexus lesion which would normally cause few or no symptoms but the presence of a CTS is somehow enhancing the symptoms and making the root/plexus lesion a problem. Rather less plausibly, the same alternative scenario could be invoked in reverse as for 4). In this case the symptoms are only those of a root/plexus lesion.
6) The patient has only a CTS
7) The patient has only a root/plexus lesion
8) The patient has neither CTS nor root/plexus problems and the symptoms have another cause entirely.
9) The patient has CTS and/or root/plexus problems but both are innocent bystanders and another disease is actually causing the symptoms

Note that these scenarios represent the underlying 'truth'. When we superimpose the uncertainties of diagnosis and testing, some patients will be miscategorised - in the simplest example, a patient with tenosynovitis may turn up test results which show EMG evidence of a root/plexus problem which simply does not exist - i.e. a simple false positive. Such a patient is in danger of being miscategorised as 7 rather than 8 if the physician is not alert to all the possibilities. Overall, I would expect an experienced clinician dealing with such cases to correctly identify and categorise in the above groups about 90% of all patients who have symptoms suggestive of either CTS or plexus/root problems.

Given the correct categorisation, patients in groups 2, and 6 will respond well to treatment of CTS. Patients in group 1 will gain relief of some of their symptoms. Patients in groups 4 and 5 may benefit from CTS treatment. Conversely, patients in groups 3, and 7 should respond well to treatment of the neck/shoulder (though this is therapeutically more difficult than treating CTS). Again patients in groups 4 and 5 may benefit from treatment directed at the shoulder too. To look at it the other way around: the groups that should NOT be treated as CTS are 3, 7, 8 and 9. The reader is left to draw his or her own conclusions as to the answer to the question at the head of this section. Groups 1, 4 and 5 could all be regarded as ‘double crush’ but group 1 are a special case because both lesions are independently causing symptoms and thus, though some of their symptoms may respond to treatment of either lesion they will gain most benefit from having BOTH treated. Such cases are probably the ones which account for failure of correctly performed carpal tunnel decompression to relieve the symptoms of a correctly diagnosed CTS in studies of failed carpal tunnel decompression. In the other cases in which 'double crush' applies (groups 4 and 5), one may only need to treat one of the two lesions to relieve all the symptoms, regardless of which is the ‘primary’ lesion of the two.

Can 'double crush' mimic the symptoms of CTS?

Underlying this question is the unspoken idea that these are two different diseases which can be mistaken for each other (see above). If we ignore this misconception for the moment the answers are: Symptoms - YES; Nerve conduction studies - only if there really is a CTS. To amplify, root/plexus disorders may produce symptoms which are very difficult to distinguish from those of CTS, though I should emphasise that I do not believe this to be common. Furthermore, it should be remembered that the symptoms of CTS may be highly atypical in a few cases. Root/plexus lesions do NOT produce the typical nerve conduction findings of CTS at the wrist though they can produce individual measurements which could be misinterpreted as CTS if seen in isolation. Similar difficulties can arise with generalised nerve disorders which makes it essential to perform multiple nerve conductions when looking for CTS. Anyone who relies on a single measurement (some people do only the motor latency from wrist to Abductor Pollicis Brevis, for example) will make mistakes.

How do 'RSI' cases fit into all this?

The bottom line here obviously relates to the frequency of occurrence of the various pathological groups outlined above in the population of patients presenting with upper limb symptoms which appear to be related to work. Some of the published evidence has suggested that high proportions lie in groups 1, 4 and 5 [1], other papers suggest that, out of all patients with ‘RSI’ the proportion of patients in groups 1, 2, 4, 5 and 6, i.e. those in whom CTS treatment may be justified, is very low and are used as evidence against surgical treatment of RSI. This same group of authors are fond of quoting the 'low success rate' of carpal tunnel surgery.

There are no absolute answers to this, even combining the wealth of published data already available but the views expressed in the previous paragraph undoubtedly represent one extreme and are, on that basis, fairly unlikely to be completely correct. A few more observations may be relevant however.

Many people presenting with work related upper limb symptoms are under the age of 40, especially those who identify themselves as RSI. In patients referred to Canterbury as ‘?CTS’ ie excluding those who had any other obvious explanation for their symptoms, there was NCS evidence of CTS in 355 out of 562 patients whose symptoms sounded plausibly like CTS (63% compared to 72% in the over 40s). These younger patients are thus less likely to have CTS and the real proportion of work related symptoms in the under 40s due to CTS will be considerably lower once we include all the other readily recognisable pathologies. I therefore suspect that at least 60% of all upper limb symptoms in this group are not due to CTS but either arise in the roots/plexus, are due to other peripheral nerve lesions, or are not nerve problems at all. Any surgical series of patients drawn from this population and treated indiscriminately for CTS will produce the 'poor success rate' which is sometimes quoted unless great care is taken to filter out only those patients with really convincing CTS for surgery. The subjective success rate for surgical decompression of CTS reported in the literature varies from 30% to 100% with an overall average of 86%

If 86% of patients gain good relief from treatment for CTS then it is unlikely that Upton and McComas's figure of 75% having significant double crush syndromes, such that diagnosis and treatment of the proximal lesion too is needed, is correct. My personal estimate would be that perhaps 30% of all patients with NCS confirmed CTS have double crush in the sense of there being some abnormality on cervical X-rays, proximal EMG studies or stretch tests but that only about 10% of these (i.e. 3% overall, roughly in line with some of the larger studies quoted above) actually have proximal disease which requires treatment in addition to treatment for the CTS in order to relieve their symptoms. Readers of this paragraph who immediately disagree should read it again and note that I am talking about percentages of patients with PROVEN CTS, not percentages of patients with RSI as a presentation - for that, refer to the previous paragraph.

Revision date - 24th February 2012