Workshop on the use of stimulation single fibre electromyography for the diagnosis of myasthenic syndromes in children held in the Institute of Child Health and Great Ormond Street Hospital for Children in London on April 24th, 2009

1. Introduction 

A meeting to discuss the use of stimulation single fibre electromyography (StimSFEMG) to assist diagnosis of congenital myasthenic syndromes (CMS) took place on 24th April 2009 at the Institute of Child Health in London. Seventeen neurophysiologists with a special interest in electromyography from 15 regional centres attended.

This workshop followed a preliminary meeting in October attended by the inventor of the technique, Professor Erik Stalberg and aimed to produce a network of neurophysiologists skilled in the technique and able to offer it as a routine investigation. It covered the technical aspects of technique, normative values, and recognition of abnormality. The summary that follows discusses the important points that emerged from the meeting.

2. Background 

The unravelling of the complex abnormalities that cause the CMS has been a rapidly expanding field of molecular genetics [1], [2], [3], [4], [5]. Unlike the situation in the autoimmune myasthenia gravis (AIMG) their clinical presentation is varied. In AIMG it is entirely reasonable to look for antibodies against the acetylcholine receptors as the first line of investigation after presentation, avoiding all contact with neurophysiology. In the CMS because of their variable clinical presentation, which includes those more commonly involved with other conditions such as hypotonia in the newborn, as one example, and the large number of possible genetic abnormalities involved, the clinicians have to have some way of determining whether it is worthwhile to start screening for the genetic defects. It is here that neurophysiology comes into its own.

There have been problems with this approach. Neurophysiologists have been unable to agree which test should be used, some are more unacceptable or indeed impossible to do in children and even if there could be a consensus the skill needed to offer these tests is not widely available. Couple this to a real difficulty in giving a numerical measure of the abnormality and it is hardly surprising that some clinicians are unaware of the need for neurophysiology in the diagnosis of CMS and indeed may not have recognised cases as a consequence.

3. Stimulation single fibre EMG as the test of choice 

The arguments for the use of stimulation single fibre EMG (StimSFEMG) as the correct test to use have already been rehearsed elsewhere [6]. In summary repetitive nerve stimulation (RNS) is only abnormal when there is a disorder of the safety factor for neuromuscular transmission sufficient to cause a block or failure of transmission and therefore will never be abnormal if the abnormality is real but less severe than that. Therefore, even if performed perfectly, itself no mean feat as the children find it uncomfortable and will move, it could still miss cases. The alternative test, single fibre EMG (SFEMG) may be abnormal if there is an increased variability of the speed of depolarisation of the post synaptic membrane, again because of alterations of the safety factor and is more sensitive. When done as originally devised it needs high levels of cooperation from the patient and high levels of skill from the operator as individual pairs of muscle fibre potentials from a single motor unit are isolated and their variability measured. This is impossible in children. StimSFEMG by stimulating the nerve itself removes the need for any cooperation.

Its acceptability to patients has been one stumbling block to the more widespread adoption of StimSFEMG as the test of choice for diagnosis of CMS. In certain part of the world, in particularly the USA, it is considered impossible to do without a general anaesthetic, which immediately restricts it widespread use as a screening test and in any case is not correct. With use of local anaesthetic cream and selection of the orbicularis oculi as the target muscle, the test can be performed routinely in children of any age without the need for significant restraint or sedation. Some centres use melatonin but not all as the test itself is so quick that it is not unrealistic to have the actual test itself, forgetting the preparation time, completed in around 5min.

4. Measures to increase the specificity of the test 

StimSFEMG has a high sensitivity and in our unpublished series of 300 cases will pick up a myasthenic disorder in around 85% of cases. Specificity can be less good. When recording stimulation single fibre it must be remembered that any disorder between the point of stimulation and the recording needle will cause an abnormality. This includes the terminal nerve, the neuromuscular junction and the muscle. Whenever an abnormality of the jitter is identified one element of the protocol was that the needle must be left in the orbicularis oculi in order to observe if there is any indication of a neurogenic abnormality. This will be seen when there is denervation and reinervation and is easily recognisable in the interference pattern. A refinement to the protocol was made that if the orbicularis oculi is not clearly abnormal it is important to make an assessment in another bulbar muscle and genioglossus is the best for this. Approached by the sub-mental route it is well tolerated and there have been instances when it has been the only muscle to show an abnormality. If found neurogenic abnormality must be sought in the peripheral muscle combined with a routine investigation of the peripheral nerves.

The distinction between a disorder of the neuromuscular junction and the muscle itself is less easy and may indeed prove not to be possible. For this reason once the neurogenic abnormalities have been ruled out the conclusion of the report should read that there is a disorder of neuromuscular transmission (NMJ or muscle) and not the neuromuscular junction alone, although overwhelmingly the NMJ will be the most likely site of the abnormality.

5. Measurement of abnormality 

Clinicians will be used to seeing figures in any report of an Electromyogram (EMG) and it is no different for studies of the neuromuscular junction. However, the figures which will be presented in this instance, the mean consecutive difference (MCD) or “jitter”, is highly unreliable in certain parts of its range and should not be included in the conclusion of the report in case this may lead to misapplication of the data. Clearly explanation for this policy is warranted. All computer programmes in the EMG machines that are commercially available have peak detection algorithms which are then used for the determination of the MCD. In volitional SFEMG most often only two peaks will be identified from two individual muscle fibres from the same motor unit. Any fluctuation in the latency of these two peaks is easily picked up and the programme gives a reliable measure of the “jitter”. As a result a fairly good linear relationship has been established between the degree of abnormality and the presumed severity of the underlying condition. In StimSFEMG because multiple motor units are being activated, large numbers of action potentials from the muscle fibres will appear, up to as many as 10 at any one time. Any slight fluctuation of the “jitter” of any of these components makes it possible that the potentials will coalesce and produce a composite wave form whose peak measurement will be both difficult and unreliable. The more abnormal the jitter the more difficult this problem becomes. An upper limit of the normal range for MCD is known and guides division of normal from abnormal but in the clearly abnormal cases the appearance of the traces is so obviously abnormal no such need exists.

The lack of a linear relationship between the MCD value and the severity of the condition is an unavoidable consequence of using StimSFEMG and is unlikely to change with refinements of the peak detection algorithms. It means that any realistic attempts to separate different types of CMS based on the range of MCD values will not occur nor will it be possible to determine whether the severity is changing should you be asked to repeat the study. The ability of the operator to determine whether the study is abnormal or not is not compromised however and the accuracy of the test’s interpretation does not suffer as a consequence.

6. Distinction neurophysiologically between different CMS 

As has already been stated the difficulties with MCD measurement bar any realistic attempts to correlate severity with CMS subgroup. That said it is part of the protocol to look for repetitive CMAPs should a neurogenic cause be excluded and this will identify those cases that are due to Slow Channel Syndrome or End-plate Acetyl Cholinesterase Deficiency; both important to recognise because of their lack of responsiveness to acetyl choline esterase inhibitors. In addition StimSFEMG lends itself nicely to the use of different stimulation rates to allow the pre-synaptic abnormalities such as Lambert Eaton Myasthenic Syndrome (LEM), which should show a definite improvement with rates over 20Hz [7], [8]. This is not yet part of the routine protocol as it may upset the children and make further collection impossible. It is a line of research that is actively being pursued.

7. Ruling out CMS 

To know that the condition that you are investigating is not CMS is sometimes as important as knowing there is an abnormality to make that diagnosis likely. We have had instances where cases have been studied with the diagnosis suspected from perhaps non-specific changes on muscle biopsy, with several genetic analyses having been undertaken but the MCD entirely normal when they eventually reach us. This is exceptionally useful for our clinical colleagues and our negative predictive value is around 90% level. One of the caveats of our group was that if we are to state that we feel able to rule out CMS a sufficient number of analyses have been made.

8. Conclusions 

All neurophysiological tests that can be used in the diagnosis of CMS have problems but StimSFEMG is the only one that is feasible to do in children. With meticulous attention to the technical details and care in interpretation of the results it offers a very real possibility of a screening test which is so needed for these rare but treatable conditions. The network that has been established has rehearsed the technical aspects of the test and has achieved an acceptable level of standardisation making results done by different operators comparable. It can also offer support to each other in the interpretation of difficulty cases. It will undoubtedly expand. It is now up to our clinical colleagues to realise this potential.

9. Participants 

	Matthew Pitt	Great Ormond Street Hospital for Children, London
	Alison Blake	Worcestershire Royal Hospital
	Arup Malik	Glasgow Children’s Hospital Yorkhill, Glasgow
	Sushma Goyal	Evelina Children’s Hospital, London
	Andrew Holton	Leeds General Infirmary
	Julian Blake	Norfolk and Norwich University Hospital
	Lesley Notghi	Birmingham Children’s Hospital
	Nandra Surjit Lyons	Poole Hospital
	Tom Tidswell	Royal Free Hospital, London
	Nofil Khalil	Charing Cross Hospital, London
	Thabit Sabbubeh	Queen’s Medical Centre, Nottingham
	Jacquie Deeb	Great Ormond Street Hospital for Children, London
	Gareth Payne	Cardiff Royal Infirmary
	Gerrard Mullins	Norfolk and Norwich University Hospital
	Andrew Michell	National Hospital for Neurology and Neurosurgery, London
	Sajjad Ali	Birmingham Children’s Hospital
	Ravi Knight	National Hospital for Neurology and Neurosurgery, London