Abstract
Introduction
The title of this paper relates to the fact that neurological examination is used by many and abused or misused by most. Two phrases are commonly found in patients’ records. The first is ‘full neurological examination was normal’; there is no such thing. It implies that no test of neurological function has been omitted. No patient is approached as if they might have any neurological condition known to medicine. This phrase usually means an arbitrary range of tests, perhaps from a checklist, and is as meaningless as a ‘full biochemical work-up’. If anybody were to attempt a full neurological examination, they would be at it all day, waste a considerable amount of time and achieve very little.
Quite often, the word ‘routine’ is substituted for ‘full’ as in ‘routine examination was normal’. This applies as much to investigations (‘routine blood tests were normal’) as it does to physical signs. Physical signs are a form of investigation and as such should be targeted. ‘Routine’ implies the mindless use of a set of tests or investigations, not targeted to the patients’ problem.
The neurological examination seems to have become a ritual rather than a set of tools. If there is a job to do, choose the right tool for that job, and use it correctly, otherwise it does not work. In medicine, it is a matter of choosing the correct sign and using it to its best effect.
Tests or signs should only be in response to a question, which might be ‘what is the haemoglobin?’ or ‘what is the plantar response?’ Of course, investigations that conform to a prepared preplanned protocol for a particular condition are entirely appropriate, such as a protocol for the investigation of a young patient with a stroke or for the investigation of a peripheral neuropathy, because in preparing such protocols, somebody has postulated the questions which the tests are designed to answer.
The history
The history is overwhelmingly the most important part of a neurological assessment. Nearly 40 years ago, Hampton et al.
1
showed that in a general medical clinic, the diagnosis could be reached on the history alone in about 80%, that examination only added about 9% and all investigations about another 9%. Although this paper was written before the widespread use of computed tomography (CT) and magnetic resonance imaging (MRI) scanners, the actual numbers might now be different. I believe that their conclusions are valid today:
Physicians can allocate the relative time spent taking the history and examining the patient with some confidence knowing that extra time spent on the history is likely to be more profitable than extra time spent on the physical examination. More emphasis must be placed on teaching students to take an accurate history and proportionally less on how to elicit physical signs. More emphasis must be placed on communication between the patient and his physician.
This is still true, particularly in neurology and about the right proportion for the allocation of time spent in a neurological consultation. Time is at a premium in the National Health Service, and none of us have as much time as we would like with each patient. The commonest cause of failure to reach a diagnosis is an inadequate history, and the commonest cause of complaints and dissatisfaction is due to a failure of communication. It is far better to spend more time on the history and explanation than to waste time on examination to no useful purpose.
The proper use of physical signs
The main use of neurological signs is in the more precise localisation of a lesion already indicated by the history and to answer the classical questions: Can this all be due to a single focal lesion? Must it be due to multiple lesions? Or is it a system disease? Rarely is a single sign by itself diagnostic of a particular disease, though a cluster of signs may be almost diagnostic.
Neurological signs should be used to clarify the history. If the diagnosis is clear from the history alone, as it often is, it might not be necessary to examine the patient at all, or perhaps look for only one or two signs. In other cases, it might be appropriate to spend a considerable amount of time over an extensive examination. There are many neurologists who go through a ‘routine’ sequence of signs regardless of the circumstances and, if challenged, justify it on the basis that patients expect it and are reassured by it, even if it adds nothing to the analysis of the problem. This is a view I do not share and one that spuriously perpetuates the perceived need for such ‘routine’ examinations. As one neurologist put it, ‘at one level, it is simply a self-serving mechanism for the perpetuation of the neurological mystique’. In a recent discussion on this topic with a distinguished neurologist, who does believe in a routine examination, he told me that if you did not do the visual fields in a patient with carpal tunnel syndrome, you might miss acromegaly. Of course, testing visual fields in somebody who might have acromegaly is absolutely correct and done because the appropriate question has been asked and not because it is No. 2 in the list of cranial nerves. Obviously, this all presupposes a cooperative patient; clearly, the whole process is modified if no history is available or the patient is unconscious, uncooperative, confused or with a conversion syndrome.
Selection of signs
So, if it is correct to be selective with the examination, what are the criteria for selection? The selection of signs should be appropriate to the problem already defined by the history. If it is done by rote, as it often is, it will always be inadequate or inappropriate. For example, testing flexor digitorum profundus to the little finger is highly relevant in ulnar nerve palsies and completely irrelevant in a patient with epilepsy. Choose signs that clearly and reliably identify pathology, such as the plantar response, or are key to the localisation of the lesion, such as an absent reflex. These are hard signs, and I find the concept of a spectrum of hard to soft signs is useful; there should be a strong bias to the use of hard signs.
The spectrum of reliability, hard to soft signs
A really hard sign clearly indicates pathology and gives lateralisation and localisation. Consider how hard or soft are the signs we use every day, which is really an assessment of how reliable a sign is in a particular circumstance, because it does depend very much on the context. An extensor planter response is a very hard sign for pathology somewhere in the upper motor neuron (UMN) pathway, but a very soft sign for localisation; it is even rather a poor sign for lateralisation, unless it is known whether the lesion is above or below the foramen magnum. A Horner’s syndrome is a superb lateralising sign but a poor localising sign. However, an absent reflex is hard on all criteria, though often irrelevant. A physical sign may be hard in one context and very soft in another, meaning quite unreliable. The abdominal reflexes are a good example, compounded by difficulty in their interpretation. There are instances in which the abdominal reflexes are important, such as looking for a lower thoracic reflex level or when present on one side and absent on the other, but this is in response to specific questions, and I suggest that the abdominal reflexes, in common with many other signs, have no value except in such very limited circumstances. Sometimes a combination of soft signs gives hard information, and nearly every sign ever described, however soft, may be useful in exceptional circumstances.
Some signs are useful if present, but useless if absent; again, weakness of flexor digitorum profundus to the little finger is a good example, and if it is affected in an ulnar nerve lesion, the lesion must be at or above the elbow, but it may be normal in lesions at the elbow. The absence of a sign may give useful information in localisation, for example, the absence of a Horner’s syndrome in lower brachial plexus palsy.
In addition to selecting hard signs, there should be a bias towards the use of motor signs, because it is common for motor deficit to be relatively under-reported and examination often reveals more signs than might be expected from the patients’ account. For example, the patient may complain of weakness and stiffness of one leg, but examination shows an asymmetrical spastic paraparesis. Conversely, cutaneous sensory symptoms are relatively over-reported compared to what is found. For example, there may be no signs detectable in a patient complaining bitterly of symptoms referable to carpal tunnel syndrome. This is just as well, because motor signs are usually easier to perform and interpret than sensory deficit, which is necessarily subjective. Furthermore, if a patient does not have any cutaneous sensory symptoms, any recently acquired cutaneous sensory deficit is very unlikely, which means that the sensory examination can be limited to testing joint position sense and vibration. There are, of course, exceptions, such as forme frustes of the Brown-Sequard syndrome and with sensory inattention; but then the tests for cutaneous sensory deficit would be in response to specific questions.
The motor system
When examining the motor system, it is worth considering the criteria which determines the movements to test; it is of course the movement, not the muscle, which is tested. Ideally, it should help in the differentiation between upper and lower motor neuron lesions; it should have a single root innervation, it should help identify an individual peripheral nerve lesion and preferably the level of the lesion. The principal muscle should be visible so that wasting or fasciculation can be seen and the muscle consistency with contraction can be felt, and some preference should be given to muscles which have an easily elicitable reflex. A good example is brachioradialis because although strength is difficult to assess, the muscle can be seen and felt, it has an easily elicitable reflex and is almost pure C6; quite important since biceps is evenly C5 and C6 so that a reduced biceps reflex and normal brachioradialis reflex is C5, whereas a reduced biceps and absent brachioradialis reflexes is C6.
Commonly tested movements.
Reproduced from ‘Aids to the examination of the peripheral nervous system’ with permission from the Guarantors of Brain. 2
It is important to test one movement at a time with the principal muscle in an optimal position, with the joint supported and with appropriate leverage. The best position of the limb, the appropriate support and amount of leverage is shown in ‘Aids to the Examination of the Peripheral Nervous System’. 2 The examination can often be quite limited, for example, if the question is: ‘is there any evidence of an UMN lesion affecting the leg?’ It is only necessary to test hip flexion and ankle dorsiflexion, if these two movements are normal and there is no weakness in that limb due to an UMN lesion. The same thing applies to the arm by testing just shoulder abduction and abduction of the index finger. These four movements are the only ones that need be tested when looking for weakness due to an intracranial lesion, with the very rare exception of lesions strictly confined to the motor cortex.
The sensory system
In sensory testing, probably the most common error is the use of too coarse a stimulus. It is easier and much more reliable progressively to coarsen a stimulus to obtain a result, so the initial stimulus should always be at or below the normal threshold. Nobody tests joint position sense at the hip first nor do they start two-point discrimination at 1 cm, but it is very common to see visual fields tested with gross stimuli, wiggling fingers or flapping arms, whereas the smallest movement possible at the distal interphalangeal joint of the index finger is easily appreciated in the peripheral field of normal subjects.
I have often seen the initial attempt to obtain a plantar response with a gouge up the middle of the sole of the foot, resulting in massive withdrawal and reluctance of the patient to have the test repeated. Whereas a gentle stroke just below the lateral malleolus may be all that is necessary.
Order of testing
Some thought should be given to the order in which signs are tested. For example, testing vision includes visual acuity, visual fields and fundoscopy. Many physicians test in this order, but it is easier to test acuity and look at the fundi before testing visual fields, because the visual fields are the most difficult and time-consuming to do. Knowing the visual acuity and whether there is any fundus abnormality makes testing and interpretation of the visual fields very much easier, since local ocular problems can be excluded. Similarly, tests of coordination should be done after tests of motor and sensory function, because it is only after the degree of weakness or sensory deficit is known, is it possible to determine whether any observed incoordination is due to a cerebellar lesion.
Useless signs
There are some signs, hallowed by long tradition, which are widely used but are almost completely useless. A good example is Rombergs test; described in 1840, which was before the neurological examination was practised as we know it today. Romberg observed that patients with tabes dorsalis and gross impairment of joint position sense fell over when they closed their eyes. This is a very crude test of proprioception and only becomes positive when joint position sense is absent to above the ankles; hardly a sensitive sign. Why not just test joint position sense at the big toe? Despite this, it is a very widely used sign and taught to all medical students; do they learn anything from it? Is swaying about normal? If so, how much swaying is allowed? Another example is the glabellar tap sign, easily demonstrated once a diagnosis of Parkinson’s disease has been made, but not a sign that makes the diagnosis.
Pursuing an abnormality
Once an abnormality has been found, it is usually necessary to undertake a series of further tests in order to define the localisation of the lesion more accurately, and this progression should be pursued until a definite conclusion is reached. A good example is in the localisation of a peripheral nerve lesion. This approach requires analysis of both the problem and the response to the tests during the examination and pursuing this until a conclusion is reached and before leaving the bedside; not performing to an arbitrary list hoping to reach a conclusion at the end, away from the bedside or after the patient has gone home. This requires cerebration during the examination, not just blindly following a check list.
Impeccable technique
Assuming an optimal choice of physical signs, the whole process may be completely negated by poor technique. The hardest sign becomes completely useless unless the technique is optimal. All tests should be applied in such a way that they detect minimal abnormalities and clearly differentiate between normal and abnormal. The answer to any question that a sign is designed to resolve is: ‘is it normal or abnormal?’ It may then be possible to undertake a qualitative assessment of the degree of abnormality, for example, into mild, moderate or severe, or use the Medical Research Council scale for muscle weakness. Quantitative assessment is seldom possible at the bedside. It is quite unacceptable to use phrases such as ‘quite good’, ‘not too bad’, ‘seems OK’, ‘relatively well preserved’ or ‘equivocal’. If a sign is thought to be equivocal, it may be because it is basically a poor quality soft sign and should not have been used in the first place or that it has been badly performed and should be done again. The commonest reason for a sign to be thought equivocal is technical incompetence.
Interpretation of signs
Many physicians attempt an analysis on inadequate information often because the examination has not been pursued to a conclusion or because a sign has been misinterpreted. This may happen with dysarthria and gait disturbance. For example, a patient’s speech or gait may be thought to have all the characteristics of a cerebellar deficit, but when examined, there is clear evidence of bilateral UMN lesions as well, a common combination in multiple sclerosis. Is there a clinician who can confidently ascribe the speech difficulty to this combination and attribute a rough proportion to each system just by listening? There is therefore no value in describing a speech disturbance as ‘scanning’, ‘staccato’, ‘slurred’, ‘ataxic’ as if these terms had a pathological connotation. It is only necessary to identify that there is a gait disturbance or dysarthria (sometimes difficult if the patient has a strong accent or where English is not the first language) without attempting further analysis, because examination will reveal the cause or causes; a vermis lesion is a rare exception.
One of the difficulties in determining whether a sign is normal or abnormal is if it is used in a way that is both difficult for the patient to perform and for the examiner to interpret. Examples of this include using too much leverage when testing muscle strength and testing more than one movement at a time; for example, testing abduction of the index finger and the little finger at the same time. This is poor technique because these two muscles are both weak in UMN lesions, are both innervated by the first thoracic root (T1) and the ulnar nerve, so no additional information is obtained by testing both muscles at the same time, and if considering a deep motor branch lesion, the two movements must be tested separately. It is difficult for the patient to concentrate on two movements at the same time, and it is difficult for the examiner to judge the strength of these two muscles, because they are of such disparate power. Much better to test first dorsal interosseous by itself; it can be tested in isolation, it can be seen and felt and it is the end of the line for the ulnar nerve. If the first dorsal interosseous is normal, there is no motor deficit due to a lesion of the main trunk of the ulnar nerve.
Testing several movements at the same time as some sort of screening test is very undesirable, because such tests are suboptimal for each movement. This is well shown in the test of opposition of the thumb and little finger, a test widely used by general physicians. This tests C8, T1, the anterior interosseous nerve, the median nerve and the ulnar nerve but does none of these things well, and no useful conclusion can be drawn from tests like this.
Another example of making a test difficult for both the patient and the examiner is moving the target about when doing the ‘finger-nose test’. In theory, this may enhances the test, but it makes the identification of a slight terminal wobble very difficult. I have tried both methods in patients with minimal cerebellar tremor and never found the tremor absent with a static target and apparent with a moving target.
Conclusion
The physical examination is only a part, and a very important part, of the neurological assessment, which must include the history, appropriate investigations and discussion, but examination will not play its part in this unless the choice of signs is appropriate and the technique impeccable.
There are fewer than 400 neurologists in the UK; yet, neurology is about 20% of general medicine, so that most neurology is done by general physicians, most neurological examinations of inpatients are done by a junior medical staff. The standard of these examinations is lamentable. Medical students and junior doctors learn their examination technique from their seniors. While it may not matter what techniques are used in the unwitnessed privacy of a consulting room, any observed examination in clinics or on the wards should demonstrate impeccable technique and the reason for the choice of signs clearly explained, especially with an audience of medical students or junior medical staff, because at their stage, it is just as easy to learn good technique as it is to pick up bad habits, always assuming that the demonstrator knows what is good technique. Judging by the standard of current practice, my generation of neurologists has failed dismally in this respect. Good neurology should not only be done but also be seen to be done, and I hope that in the future, we will see less irrelevant, inaccurate, inappropriate and inept neurological examinations, though with the current reliance on diagnostic technologies (CT – computerised tomography; MRI – magnetic resonance imaging; PET – positron emission tomography; EEG – electroencephalography; EMG – electromyography, etc.) this may be a forlorn hope.
There is no ‘full’ or ‘routine’ neurological examination. History is by far the most important part of a neurological assessment. Examination can often be quite limited; signs should be used to clarify the history. There should be a bias towards the use of motor signs. All tests should be applied in such as a way that they detect minimal abnormalities and clearly differentiate between normal and abnormal. Pursue an abnormality until a clear conclusion is reached. Consider how reliable (hard to soft) a sign is in context; some signs are widely used but are almost completely useless. Good neurology should not only be done but also be seen to be done.