Pract Neurol 6:298-307 doi:10.1136/jnnp.2006.101261
  • Review

The trouble with neuralgic amyotrophy

  1. Nens van Alfen
  1. Neurologist/Clinical, Neurophysiologist Department of Neurology and Clinical, Neurophysiology Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands
  1. Correspondence to:
 Dr N van Alfen, 920 KNF, Radboud University, Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, the Netherlands;


    Imagine yourself waking up early one morning with a severe pain in your shoulder and upper arm. At first you attribute this to some strain a day or so before, but after a few hours the pain has become so severe that you go to your family practitioner or emergency department in a slight panic. The doctor looks at your shoulder and arm but sees nothing amiss, apart from the fact that you are constantly holding your arm close to your body, keeping it immobile. You are reassured that nothing appears to be seriously wrong, given a non-steroidal anti-inflammatory drug (NSAID), and sent home with advice to see a physical therapist. After a few sleepless weeks because of the relentless pain, and several unsuccesful therapeutic sessions, the therapist notices that your scapula seems to be protruding and the whole shoulder looks as if it has sagged. Some muscles are atrophic too. You are therefore referred to a neurologist or orthopaedic surgeon because of suspected radicular or shoulder joint pathology. Hopefully there will not be too much delay spent on a waiting list, and if you’re lucky you will be one of the 40% where the consulting specialist recognises the clinical picture. The chances are however that you will have to wait again for a radiological study of the cervical spine or shoulder region, and because this usually is normal you are either sent back to your family doctor or physical therapist, or referred for a second opinion, still without a proper explanation for your symptoms. Eventually–but this may often be more than six months later—someone thinks of a peculiar neurological disorder called neuralgic amyotrophy, and you find yourself discharged from further specialist care with the message that nothing can be done and spontaneous recovery will follow in a few years—which does not always happen.

    If this story sounds somewhat exaggerated, rest assured, it is not. It is a very typical referral pattern for many neuralgic amyotrophy patients. For most neurologists, neuralgic amyotrophy (or Parsonage-Turner syndrome) is likely to be one of those spot-diagnosis disorders, occasionally encountered in the form of a winged scapula. But although 70% of these patients do have serratus weakness, neuralgic amyotrophy is much more than that, and it is probably more frequent than we imagine.

    The disorder has several public relations disadvantages:

    • it is often not considered as a diagnosis for acute shoulder and arm pain because the doctor does not know about it;

    • it is often discarded as a diagnosis because the clinical picture is deemed “atypical”; and

    • patients suffering from it are often told that all will be well in a few years and then left to themselves.

    In my 10 year experience with several hundred patients, there is no reason why things should be this way. Neuralgic amyotrophy has a well delineated, if broad, phenotype that can usually be recognised just by taking a history, supplemented with a focused neurological examination.1 One only has to know what to ask and look for. This review provides the information required to answer two clinical questions: “Could this patient be suffering from neuralgic amyotrophy?” and “This patient has neuralgic amyotrophy, so what to do next?” I will elaborate on the most common clinical findings, pitfalls, and differential diagnosis and provide a practice based (if not formally evidence based) overview of therapeutic strategies.


    Clinical features and facts

    Once seen or heard about, neuralgic amyotrophy presents as a very distinct clinical syndrome 95% of the time. The core features are shown in table 1.14 The characteristic pain and subsequent patchy motor and sensory findings are the keys to the diagnosis. Their typical features are described below.

    TABLE 1

    Neuralgic amyotrophy at a glance

    Acute pain

    Acute onset of pain in the neck, shoulder, upper arm, and/or forearm region occurs in 96% of all patients (fig 1). It wakes 61% of patients in the early morning at onset, and usually increases to maximum severity in a few hours. It is typically relentless and unlike any pain experienced before (unless there was a previous attack). The severity is best illustrated by instructing the patient to grade the pain on a scale of 0 to 10 (0 = no pain, 10 = the most severe pain imaginable). Initially 60% and subsequently 90% of the patients have a score of 7 or more; the median score is 8 at onset and 9 at maximum intensity. Pain is usually worse at night and disturbs sleep in 94%. The pain can sometimes have a stuttering onset; in 20% of the attacks it is present on and off for several weeks before becoming continuous. It lasts for about four weeks on average, and is usually longer in men with a mean duration of 45 days versus 23 days in women. In 5% of cases the pain disappears within 24 hours, in 23% it lasts for 1–7 days, and in 10% it lasts for 60 days or more. Typically acetaminophen (paracetamol) or NSAIDs provide no relief whatsoever. Pain often causes patients to sit up all night with their arm pressed tight against their body, or has them crawling desperately on the floor not knowing what to do anymore. This story is not easily forgotten once heard, and the next patient will probably tell it too, making it a useful diagnostic tool.

    Figure 1

    “Typical acute stage neuralgic amyotrophy” (with kind permission of the artist and her father).

    After the acute stage of an attack 77% of the patients will go on to suffer from additional pain types

    Subsequent pains

    After the acute stage of an attack 77% of the patients will go on to suffer from additional pain types:

    • Following the initial continuous pain, normal arm posturing or movements can elicit shooting or radiating pains by mechanical distortion of damaged plexus nerves, which resembles the pain evoked by leg or arm raising in radiculopathies. This hypersensitivity usually dissipates in weeks to months.

    • Additionally 65% develop musculoskeletal pain at the origin or insertion of the paretic or compensating muscles, especially in the periscapular, cervical and occipital regions. This pain is often chronic, therapy resistant, and can become more incapacitating than the residual paresis. The main cause of musculoskeletal pain is scapular instability due to paresis of the serratus anterior (or trapezius). Myalgia is often found in the serratus region, from the medial scapular edge, subscapular region, and axilla to the mid-axillary chest wall where the muscle inserts onto the ribs. Because chest wall movements increase this pain, it can even prevent normal breathing, and be particularly disabling if there is also phrenic nerve dysfunction. As well as in the serratus region, this pain can occur in the area of any paretic muscle.

    • Compensating muscles that become painful are mainly the trapezius, levator scapulae, and pectoralis minor, because of strain in their attempt to overcome the downward and lateral scapular displacement (fig 2).

    • Additionally, pain can result from glenohumeral joint pathology such as rotator cuff impingement or capsular irritation, developing as a consequence of the altered shoulder biomechanics.5

    Figure 2

    Downward displacement and rotated position of the scapula caused by serratus anterior paresis. (A) Mild displacement on the right, normal scapular position on the left. (B) More severe displacement with compensatory trapezius tensing on the right, also mild displacement on the left (notice that in both cases there is no significant winging at rest).

    Motor and sensory symptoms

    Another typical feature is the patchiness of the motor and sensory symptoms and signs. This can make it difficult for clinicians to recognise the disorder. Sometimes an attack consists just of a few hours of annoying pain in the arm and loss of pinch grip for a few months, caused by an anterior interosseus nerve lesion. Or the patient may have severe pains in both shoulders and arms for weeks on end, orthopnoea requiring nocturnal positive pressure ventilation, and little functional use of the upper extremities for several years because of severe bilateral plexus damage including bilateral phrenic nerve palsy. Any part of the brachial plexus, and so any muscle or skin area supplied by the plexus can be involved, in all sorts of combinations. Localisation of the sensory symptoms and the pain tend not to correlate with the paresis. It is precisely the recognition of this patchiness that is a very important clue to the diagnosis of the plexopathy in neuralgic amyotrophy.

    However, there are several patterns of paresis that occur more frequently than others:

    • Most common (71%) is an upper brachial plexus distribution—the periscapular and perihumeral muscles—with (50%) or without (21%) involvement of the long thoracic nerve (fig 3). Overall, serratus anterior and infraspinatus paresis occur in 70% of the attacks, biceps in 60%, deltoid in 50%, trapezius in 20%, and pectoralis major in only 15%.

    • Women have symptoms in a middle and/or lower brachial plexus distribution twice as often (23%) as men (11%), including anterior interosseus nerve dysfunction with reduced pinch grip because of weakness of the long thumb and index finger flexors.

    • Almost 80% of the patients have sensory deficits. Moreover, neuralgic amyotrophy sometimes, despite its name, presents with pain and sensory symptoms only, frequently in the distribution of the lateral cutaneous nerve of the forearm.6 In 15% of the attacks there is distal sympathetic dysfunction such as hand oedema and vasomotor instability.

    • Nerves outside the distribution of the brachial plexus are involved in 56% of the attacks in hereditary neuralgic amyotrophy and in 17% of non-hereditary cases, usually in combination with some degree of brachial plexus involvement. For example, the lumbosacral plexus (usually lumbar trunk), phrenic nerve, and recurrent laryngeal nerve can be involved. Symptoms from such involvement can also occur as an isolated clinical syndrome without brachial plexus involvement. Occasionally other nerves such as the facial or abdominal intercostals are affected. Phrenic nerve palsy occurs in 7% of idiopathic cases and 14% of hereditary patients, and presents as orthopnoea or dyspnoea on stooping, dyspnoea on effort, and sleep disturbances (especially during REM sleep when only the diaphragm is active).

    • Bilateral brachial plexus involvement occurs in almost 30% of patients and is usually asymmetric and patchy. In attacks with more extensive involvement of the brachial plexus(-es), symptoms and deficits in the leg, phrenic, or other nerves are more commonly found.

    • The phenotype is limited however: a perfectly symmetrical complete di- or tetraplegia is not seen, nor is a Horner’s syndrome as a rule. If present these features should prompt a search for an alternative diagnosis.

    Figure 3

    The distribution pattern of classic neuralgic amyotrophy within the plexus (in yellow) and on physical examination (in red). Musculocut, musculocutaneous nerve; CAL, lateral cutaneous nerve of the forearm; PT, pronator teres muscle; SA, serratus anterior muscle; SS, supraspinatus muscle.

    Figure 4

    The patient adducts and elevates the right shoulder blade by tensing his trapezius muscle, to compensate for scapular instability. Simultaneously a glenohumeral subluxation is visible, caused by a combination of scapular displacement and deltoid paresis.

    Bilateral brachial plexus involvement occurs in almost 30% of patients

    Physical examination

    Because neuralgic amyotrophy attacks frequently involve the proximal or upper parts of the brachial plexus and so influences shoulder biomechanics,7 there are some issues that merit special attention during the physical examination. As an example, a routine I have found helpful in daily practice is shown in table 2.

    TABLE 2

    The physical examination in neuralgic amyotrophy

    A pitfall in the diagnostic phase is that patients tend to complain about that part of their shoulder or arm which has the maximum pain or paresis, but pay hardly any attention to lesser impairments in strength or sensory loss. The asymmetric involvement in bilateral attacks can for example be easily overlooked. As a clinician one should be aware of this, and supplement the history with a meticulous physical examination that specifically pays attention to those muscles or skin areas the patient doesn’t mention. Omitting this increases the chance of diagnostic errors (for example, making the clinical diagnosis of a mononeuropathy or radiculopathy instead of a plexopathy), and errors in correctly explaining the symptoms, such as mistaking a serratus paresis with scapular displacement for trapezius atrophy, or inability to abduct the arm for a deltoid paresis.

    Diagnosing neuralgic amyotrophy

    Acute phase

    Patients are encountered in either the acute or chronic phase of the disorder. In the acute phase pain usually dominates the clinical picture, and the challenge lies in a swift identification or exclusion of several possible causes for such pain, and assessing features that make neuralgic amyotrophy more—or less—likely.

    A simple three-step way to differentiate the disorder from its most common differential diagnoses is to start with the following questions:

    • Is the pain acute, very severe (score >7) and unlike anything the patient has had before? If yes, neuralgic amyotrophy is likely; if no, it is possible but one should consider alternatives.

    • Is there limitation of passive arm external rotation or abduction? If no, neuralgic amyotrophy is likely; if yes, shoulder joint pathology (bursitis, tendinitis calcarea) is more likely.

    • Are all the symptoms (pain, paresis, and sensory disturbances) in the same root distribution? If no, neuralgic amyotrophy is likely; if yes, a cervical radiculopathy is more likely.

    Table 3 shows the differential diagnosis of upper extremity pain and/or paresis, and table 4 the differential diagnosis of plexopathies.8 The lists are not meant to be exhaustive, but can be used as a guide to further examination.

    TABLE 3

    Differential diagnosis of upper extremity pain and/or paresis

    TABLE 4

    Differential diagnosis of the causes of plexopathy

    Chronic phase

    Encountering a patient in the chronic phase, months or even years after onset, makes for a completely different diagnostic challenge. These patients are often referred for complaints of painless weakness or chronic cervicobrachial pain, with the (unexpected) finding of focal atrophy or scapular displacement indicating that “something neurological must be wrong”. By this time any pain is often musculoskeletal rather than neuropathic. This may confuse the practitioner because pain and atrophy do not appear to stem from the same cause. However, from the above it will be clear that both can follow patchy upper extremity axonal loss and thus altered biomechanics.

    Role of EMG and other laboratory investigations in diagnosis

    Routine laboratory tests do not contribute to the diagnosis, but can be useful to rule out rare direct infectious causes such as HIV or Borrelia in patients at risk. An EMG study is often used to confirm the diagnosis by showing an axonal plexopathy and to exclude a monoradiculopathy, entrapment neuropathy, or conduction block. However, after a proper history and examination the diagnosis is usually clinically obvious, and I mainly use the EMG to assess the severity of the axonal damage, especially in the functionally important muscle groups. This helps to predict the prognosis and can reassure the patient that recovery is actually taking place even though it has not yet resulted in any functional improvement.

    Many patients will have MRI imaging of the cervical spine. In a large cohort abnormalities such as degenerative disk disease were found in over 50% of neuralgic amyotrophy patients, but these findings could neither explain the clinical picture nor course. This is a warning to place the emphasis on the clinical rather than the radiological picture.1 Plexus MRI is typically normal except for T2 hyperintensities in 5%. Chest x ray is often performed to exclude a Pancoast tumour—the course and clinical picture rarely mimic neuralgic amyotrophy, but it is most useful to document or reveal a (hemi-) diaphragmatic palsy.


    This patient only has pain, can that be neuralgic amyotrophy? Pain usually precedes paresis by one to two weeks, but in 25% it takes more than two weeks to develop a clinical deficit. This makes it difficult to reach a definite diagnosis during that early stage. In these cases all one can do is provide adequate pain relief in combination with timely follow up, and instruct the patient to return if, and as soon as, any weakness appears, so treatment (corticosteroids, see below) may be started.

    This patient seems to have neuralgic amyotrophy but the pain persists. Shouldn’t I consider other diagnoses?

    A frequent misconception is that the pain should be gone in two weeks. The average duration of the acute phase pain is 3–4 weeks, but the main point is that “neuralgic amyotrophy pain” consists of at least three different pain patterns appearing in the course of the disorder (see above). Understanding the nature of the current pain and what affects it, in combination with the assessment of muscle strain, nerve hypersensitivity, and shoulder joint pathology usually explains why it persists. Only when these features do not fit the neuralgic amyotrophy patterns should an alternative diagnosis be considered.

    This patient has fasciculations, doesn’t that suggest motor neuron disease instead of a plexopathy?

    Fasciculations are not uncommon in muscles affected by neuralgic amyotrophy. They occur either at the onset of weakness (probably a form of denervation hypersensitivity) and in the chronic phase, especially when axonal loss has been severe.

    The EMG was normal, so a plexopathy is ruled out

    Because the clinical deficits are so patchy, and often nerves are involved that lead to weakness in muscles not commonly examined during needle EMG, it is actually quite easy to miss the relevant abnormalities. For instance, the fact that the biceps brachii muscle is normal does not mean that the brachialis will be too, and a normal infraspinatus doesn’t exclude involvement of other C5-C6/upper plexus trunk muscles. So, an efficient and helpful EMG has to be based on a thorough clinical examination which points to which muscles to study. Without it, a “negative” EMG has no meaning in this context (unless each and every muscle in the affected extremity is studied—possible, but unlikely in daily practice).

    Routine laboratory tests do not contribute to the diagnosis

    Sensory deficits make neuralgic amyotrophy unlikely

    Sensory symptoms are actually quite common: 70% of the patients notice them, and in 80% they are found on examination. Many patients experience paraesthesia at onset. Deficits consist of patchy areas of cutaneous hypoaesthesia, but when they occur in the shoulder or upper arm region they often go unnoticed, the most likely explanation being that people don’t “feel” with their shoulders. But once the middle or lower trunk nerves are involved, with sensory deficits in the forearm or hands, patients will usually complain about them. Proprioception disturbances are rarely, if ever, seen.

    All patients recover after a few years

    Although this is still what some neurological textbooks say, it is simply not true: almost two thirds of the patients are left with residual pain and/or paresis even after three years or more.1 In particular, pain, altered shoulder biomechanics, persisting wrist and finger extensor paresis, or diaphragm dysfunction make recovery less than satisfactory for many patients.


    Pharmacological treatment

    Pain in the acute phase responds best to a combination of a long acting NSAID and an opiate, such as diclofenac slow release 100 mg bd with slow release morphine 10–30 mg bd. With a pain score >7 anything less is usually not helpful and should not be used. Analgesics for neuropathic pain such as gabapentin, carbamazepine, or amitriptyline may be helpful for the later stretch-sensitive nerve pain. They are not really suited for use in the acute phase because of their delayed effect, which usually takes as long to set in (several weeks) as the acute phase pain lasts.

    Almost two thirds of the patients are left with residual pain and/or paresis even after three years or more

    Preliminary data suggest that corticosteroids in the acute phase may favourably influence the course of the symptoms in some patients, but no formal trials have been conducted, and evidence is at the expert opinion level at best.1

    In my personal experience some patients do seem to benefit from a two week course of high dose oral prednisone, but only when it is given promptly. This is probably because steroids can halt inflammation within the plexus but cannot undo any damage already done. A randomised placebo controlled trial of oral prednisone is currently underway.

    Physical therapy and rehabilitation

    Treatment of the chronic musculoskeletal pain requires helping the patient re-establish optimal biomechanics. This is done by physical therapy to maintain fluent periscapular motion, preventing dysfunctional compensatory strategies, and well timed alternation between daily activities and rest. Overall, patients who keep their affected limb active tend to be better off functionally. There are some pitfalls however when it comes to exercise. Therapy failure often occurs when attempts are made to strengthen the intact compensatory periscapular muscles while the serratus anterior is weak. Although the concept seems valid, there are almost no exercises of this type that can be performed without concomitant serratus activation. This will lead to additional myalgia and fatigue in this muscle which will in turn thwart further training attempts. Another pitfall is training muscles with strength below MRC grade 3 using weights. Here one can safely assume that if a muscle is not able to lift the attached bodyweight against gravity it will certainly not be able to do so if extra kilos are added. In general, trying to train muscles that are already strained due to paresis or altered biomechanics will inevitably risk additional strain, with increased pain and fatigue.

    A foam rubber sling to relieve the scapula-stabilising muscles of the weight of the upper limb can be useful if the serratus is weak. Quite often patients already apply this support “trick” by tucking their arm and hand in a coat pocket when sitting, standing, or walking. But simultaneously the patient should maintain full joint motion a few times a day by passive or active movement to prevent contractures.

    Commercially available scapular braces also provide support, but to fixate the scapula against the thorax for maximum stability they have to be worn very tightly around the chest. This is uncomfortable, may restrict normal breathing, and prevents arm abduction more than 110° because the necessary scapular movement is limited. Patients who have found them helpful usually have worn them less tightly, but just enough to provide a bit of extra of support for the arm and improve scapular position on the chest wall.

    Patients with phrenic nerve palsy and orthopnoea should be evaluated for nocturnal hypoventilation and REM-sleep apnoeas, and treated with positive pressure ventilation if necessary.

    Treatment of glenohumeral joint pathology involves physical therapy and local injections when capsular or tendon/cuff irritation or bursitis are present. In the case of a partial tendon cuff rupture, rest is advised; a complete rupture has to be surgically treated. It is also important to look for work or daily life associated movement patterns that sustain this complication (especially repetitive arm movements or excessive weight bearing) and make the patient aware of this association.

    Since initially the severe pain will prevent many patients from going to work, the inevitable question will arise of when is it wise to return to work. The right answer is “as soon as possible”, because there is no evidence that physical activities will prevent nerve recovery or provoke another attack. However, it’s obvious that depending on the type of posture and the activities required, certain jobs increase the chance of complications. A gradual increase in activities, limiting the amount of time spent in certain postures (for example, sitting behind a desk, working with a keyboard, lifting objects), and alternating these physically demanding activities with other less demanding tasks will usually be preferable as long as weakness and pain are significant.

    Fatigue occurs in over a quarter of the patients after a neuralgic amyotrophy attack and thus seems to be part of the phenotype. Spreading one’s activities over the day and choosing what to spend one’s energy on seems the most reasonable advice for these patients. For more specific help and guidance we offer patients consultation with a rehabilitation physician as part of standard care.


    • Neuralgic amyotrophy is a distinct clinical syndrome with a broad but well delineated phenotype that is probably underrecognised.

    • The characteristic episodes or “attacks” with initial severe neuropathic pain and subsequent patchy motor and sensory findings in a brachial plexus distribution are the key to the diagnosis.

    • In one third of patients the symptoms are bilateral, but asymmetrical, and about 10% also have involvement of the phrenic nerve or lumbosacral plexus during an attack.

    • The initial pain lasts for about four weeks on average and is often followed by two other pain types that can persist for months to years; altered shoulder biomechanics due to scapular paresis and instability are the most frequent causes of persisting pain and functional deficits.

    • Recovery is not as optimistic as often stated in the textbooks; over half of the patients suffer from persisting pain and paresis even after three years.

    • Adequate analgesia including the use of opiates in the acute phase, and appropriate physical therapy and rehabilitation in the chronic phase, are the cornerstones of management.


    This article was reviewed by Dr Michael Donaghy, Oxford, UK.