OUP user menu

Phenotypic variants of autoimmune peripheral nerve hyperexcitability

Ian K. Hart, Paul Maddison, John Newsom‐Davis, Angela Vincent, Kerry R. Mills
DOI: http://dx.doi.org/10.1093/brain/awf178 1887-1895 First published online: 1 August 2002

Summary

Clinicians use many terms including undulating myokymia, neuromyotonia, Isaacs’ syndrome and Cramp–Fasciculation Syndrome to describe the motor manifestations of generalized peripheral nerve hyperexcitability (PNH). Our previous findings in a selected group of patients with undulating myokymia or neuromyotonia, and EMG doublet or multiplet (‘myokymic’) motor unit discharges, indicated that an autoantibody‐mediated potassium channelopathy was likely to be the cause of their disorder. This prompted us to search for a common pathogenesis in a wider spectrum of PNH syndromes. We studied the clinical, autoimmune and electrophysiological features of 60 patients presenting with acquired PNH. Patients were grouped according to an EMG criterion: the presence (group A, n = 42) or absence (group B, n = 18) of doublet or multiplet myokymic motor unit discharges. The average ages of onset in the two groups were 45 and 48 years respectively. The relative frequency and topography of the clinical features were similar in both groups. Serum voltage‐gated potassium channel (VGKC) antibodies were detected using a 125I‐α‐dendrotoxin immunoprecipitation assay in 38% of group A and in 28% of group B. Autoimmune disease and other autoantibodies were present in both groups more frequently than would be expected by chance (59 and 28%, respectively)—particularly myasthenia gravis and acetylcholine receptor (AChR) antibodies. The neurological disorder in both groups could occur as a paraneoplastic condition. Thymoma was detected in 19 and 11%, respectively, and lung cancer in 10 and 6%, respectively. An axonal neuropathy was present in six (14%) of group A and in one (6%) of group B patients. Thus, despite the discrete EMG distinction, both groups share clinical features often associated with autoimmune‐related diseases, which can occur as paraneoplastic disorders and, importantly, have an increased frequency of VGKC antibodies. We conclude that autoimmunity, and specifically VGKC antibodies in many cases, are strongly implicated in the pathogenesis of both groups, and that the EMG features reflect quantitative rather than qualitative differences between the diverse clinical syndromes. These findings also have relevance for disease management. A classification is proposed that distinguishes immune‐mediated PNH (irrespective of whether VGKC antibodies are detectable by standard assays) from non‐immune forms of PNH that include toxins, anterior horn cell degeneration in motor neurone disease and genetic disorders.

  • Keywords: neuromyotonia; Cramp–Fasciculation Syndrome; myokymia; peripheral nerve hyperexcitability; potassium channelopathy; paraneoplastic syndrome
  • Abbreviations: AChR = acetylcholine receptor; PNH = peripheral nerve hyperexcitability; VGKC = voltage‐gated potassium channel

Introduction

Generalized peripheral nerve hyperexcitability (PNH) usually presents as spontaneous and continuous muscle overactivity. The clinical features of the motor nerve dysfunction are diverse and include cramps, muscle twitching (fasciculations or myokymia), stiffness, pseudomyotonia (delayed muscle relaxation after contraction) and pseudotetany (spontaneous or evoked carpal or pedal spasm, or Chvostek’s sign) (Isaacs, 1961; Mertens and Zschocke, 1965; Tahmoush et al., 1991). Some patients experience paraesthesias and numbness implying sensory nerve involvement, or hyperhydrosis that might represent autonomic dysfunction or be secondary to muscle overactivity. There may also be associated CNS disorders such as mood change, sleep disorder or hallucinations (reviewed in Newsom‐Davis and Mills, 1993; Serratrice and Azulay, 1994; Barber et al., 2000; Liguori et al., 2001).

Denny‐Brown and Foley (1948) were the first to characterize the clinical and electrophysiological motor phenomena of PNH—describing the disorder as undulating myokymia. Since then, clinicians have proposed a variety of terms to describe combinations of these symptoms and signs, usually based on the dominant clinical or electrophysiological motor feature in a small series of patients. These have included continuous muscle fibre activity, quantal squander, Armadillo syndrome, Isaacs’ syndrome, neuromyotonia, Mertens’ syndrome, Mertens–Isaacs’ syndrome, pseudo‐ or normocalcaemic tetany, neurotonia, continuous motor nerve discharges, generalized myokymia and idiopathic generalized myokymia (reviewed by Hart et al., 1999). Moreover, those patients with prominent cramps may be diagnosed as having either the muscular pain–fasciculation (Hudson et al., 1978) or Cramp–Fasciculation Syndrome (Tahmoush et al., 1991). None of these terms reflects the potentially diverse pathogenesis for the nerve hyperexcitability. In some patients, for example, nerve hyperexcitability is hereditary and usually associated with various types of motor polyneuropathy. However, in the great majority, it is acquired.

We previously showed that patients exhibiting ‘myokymic’ EMG discharges (i.e. doublet, triplet or multiplet motor unit discharges), which we designated ‘acquired neuromyotonia’, appeared to have an autoantibody‐mediated nerve potassium channelopathy (Sinha et al., 1991; Newsom‐Davis and Mills, 1993; Shillito et al., 1995; Hart et al., 1997). This led us in the present study to investigate 60 of our patients, each of whom had clinical features of acquired spontaneous muscle overactivity (muscle twitching and/or cramps). We divided the patients into two groups defined by a single objective criterion: the presence (group A) or absence (group B) of myokymic EMG discharges. Our aim was to determine whether there were other clinical or electrophysiological features that clearly distinguished them and, in particular, whether the evidence for autoimmunity in the former group extended to the latter.

Methods

Patients

We studied all patients presenting between 1990 and 2000 to our neuromuscular clinics in Oxford (1990–2000) or in Liverpool (1996–2000) with symptoms or signs of acquired spontaneous muscle overactivity (muscle twitching or muscle cramps) affecting at least two regions of skeletal muscle other than the calves. Follow‐up was between 1 and 11 years. No patient had a history of exposure to toxins, a family history of parental consanguinity or primary neurologic disease (including hereditary neuropathy or episodic ataxia). We excluded only two patients from our analysis, who had developed motor neurone disease.

Patients were separated into two groups determined by the presence (group A; n = 42) or absence (group B; n =  18) of doublet, triplet or multiplet (‘myokymic’) EMG discharges.

Evaluation

Each patient was fully evaluated clinically by one of us (J.N‐D or I.K.H.). All patients had one or more EMG examinations with needle or surface recordings made for at least 5 min from each of two to five historically or visibly overactive muscles. Most patients on medication for symptomatic treatment of muscle overactivity (usually carbamazepine) stopped therapy at least 24 h before the electrophysiological tests. Similarly, pyridostigmine was stopped the day before recording in the single myasthenia patient who was receiving this medication. All patients had nerve conduction studies.

Serum was assayed for antibodies to voltage‐gated potassium channels (VGKCs; see below for method), acetylcholine receptors (AChRs), antinuclear antigen, striated muscle, thyroid, rheumatoid factor, reticulin, glutamic acid decarboxylase and double‐stranded DNA.

Creatine kinase, thyroxine and thyroid‐stimulating hormone were measured in most patients. Many patients had a CSF examination. Thoracic CT scans were performed when indicated clinically.

Voltage‐gated potassium channel antibody assay

This immunoprecipitation assay used 125I‐labelled α‐dendrotoxin as a VGKC‐specific ligand and was adapted from a method performed previously by Hart et al. (1997). Briefly, protein was extracted from rabbit brain and used as a source of VGKC. Aliquots of brain extract were incubated with a saturating concentration of 125I‐α‐dendrotoxin (1 nM). Five microlitres of patient serum diluted 1 : 10 in 20 nM phosphate buffer, pH 7.4, and 0.1% Triton X‐100 (PTX buffer) was added to 50 µl brain extract. After incubation for 1 h at room temperature, 100 µl of goat anti‐human IgG antibody (diluted 1 : 1 in PTX) was added. One millilitre of PTX was added after a precipitate had formed and the sample was centrifuged at 13 000 r.p.m. for 3 min. The pellets were washed twice in PTX and counted in a gamma counter (Canberra Packard). Results were expressed as picomoles of 125I‐α‐dendrotoxin‐binding sites precipitated per litre of serum (pM) after subtraction of the mean result of sera from three healthy individuals. Each serum sample was assayed at least twice. Antibody titres were considered positive if >100 pM, (the mean titre of a group of control sera plus 3 SD).

Results

Epidemiology

Men outnumbered women in both groups (1.8: 1 and 2.6: 1 in groups A and B, respectively; see Fig. 1A and B). The average age of onset was similar (45.8 years and 47.7 years, respectively), with broadly uni‐modal distributions that peaked in the fifth decade. Five patients in group A presented under 30 years of age, the youngest being 9 years old. No patient in group B developed symptoms before the age of 30 years. Average time to diagnosis was 4.4 years (range 0.25–30 years) and 3.0 years (range 0.25–9 years), respectively.

Fig. 1 Distribution of the ages of onset of group A and group B patients.

Clinical features

The presenting clinical features are listed in Table 1 and ranked according to their frequency—the order proving to be identical in the two groups. By definition, both groups had muscle twitching and/or muscle cramps. The proportion affected by each of the other associated motor features was broadly similar in the two groups.

View this table:
Table 1

Presenting features

Symptom/signGroup A (n = 42)Group B (n = 18)
n (%)n (%)
Muscle twitching42 (100) 17 (94)
 Symptom4016
 Sign3412
Muscle cramps37 (88)18 (100)
 Symptom3718
 Sign3113
Muscle stiffness29 (69)13 (72)
Hyperhydrosis 23 (55)6 (33)
Muscle weakness15 (36)5 (28)
Pseudomyotonia15 (36)4 (22)
Sensory symptoms14 (33)7 (39)
CNS symptoms12 (29)4 (22)
Muscle hypertrophy9 (21)4 (22)
Pseudotetany8 (19)1 (6)
Muscle wasting1 (2)0
Autonomic symptoms1 (2)0
Exercise triggered34 (81)13 (72)

When present, pseudomyotonia (delayed muscle relaxation secondary to PNH) always affected handgrip. Only three patients had percussion pseudomyotonia. Pseudotetany (spontaneous carpal or pedal spasm) was uncommon in both groups. Only two patients had a positive Trousseau’s sign, one of these having a positive Chvostek’s sign. The non‐fatiguable muscle weakness observed in about one‐third of patients was attributable to PNH rather than to myasthenia gravis or a demyelinating or axonal neuropathy. The most overactive muscles tended to be the weakest. Exercise or muscle contraction triggered or exacerbated all motor features in the great majority of patients. Tendon reflexes were normal, except in those with a demyelinating or axonal polyneuropathy.

Sensory manifestations, usually distal limb paraesthesias or dysaesthesia and numbness, were reported by about a third of patients in both groups (Table 1). Many of them had positive sensory phenomena without numbness and with normal nerve conduction, suggesting that peripheral sensory as well as motor nerve hyperexcitability can occur in the absence of an axonal or demyelinating peripheral neuropathy (Lance et al., 1979).

Hyperhydrosis was more frequent in group A patients (Table 1). While this may have been secondary to an increased basal metabolic rate caused by increased muscle activity (Isaacs, 1961), direct autonomic involvement would be a plausible alternative explanation. Indeed, one patient in group A reported symptoms other than hyperhydrosis that suggested primary autonomic involvement (Halbach et al., 1987). This patient had urinary hesitancy, abdominal bloating, constipation, and impotence that fluctuated in parallel with his muscle overactivity, although autonomic function testing in the clinic was normal.

Most group A patients reported that their muscle overactivity caused mild to moderate incapacity that interfered with their activities of daily living. One patient in this group with onset in childhood, whose symptoms were untreated for many years, had growth retardation and presented in a wheelchair with severe muscle stiffness and cramps that had been mistaken for spasticity. By contrast, most group B patients found their symptoms inconvenient rather than disabling. All patients reported that their symptoms fluctuated in severity over periods of months. No patient had a spontaneous remission.

Topography of motor features

At presentation, ∼70% of patients in both groups had limb and trunk or limb muscle overactivity (Table 2)—the regional patterns of muscle involvement being remarkably similar. The major difference between the two groups was bulbar involvement in four group A patients who did not have myasthenia gravis. Patients in both groups had facial muscle overactivity, which usually presented as twitching. This finding further emphasizes that the lower cranial as well as peripheral motor nerves can be hyperexcitable.

View this table:
Table 2

Topography of muscle overactivity

Muscles affectedGroup A (n = 42)Group B (n = 18)
n (%)n (%)
Limbs and trunk17 (40)7 (39)
Limbs only12 (29)6 (33)
Limbs, trunk + face7 (17)3 (17)
Limbs, trunk, face + bulbar 4 (10)0
Limbs + face1 (2)2 (11)
Limbs + bulbar 1 (2)0

Muscle overactivity affected lower limb more than upper limb muscles, suggesting that the nerve hyperexcitability was length dependent; the longer nerves being worst affected. Although many patients had truncal muscle overactivity, only one patient (in group A) had dyspnoea caused by chest muscle stiffness.

Associated immune‐related diseases

Autoimmune disorders occurred more frequently than would be expected by chance in both groups. Myasthenia gravis was present in 21% of group A and 12% of group B (Table 3). Manifestations of muscle overactivity appeared at the same time as myasthenia in six out of 11 patients and between 4 to 34 years after the onset of myasthenia in the other five patients. Most myasthenia patients in both groups had a thymoma. Three patients had a thymoma and AChR antibodies, but no clinical or EMG evidence of myasthenia. In three patients, evidence of motor nerve hyperexcitability was present for 1–3 years before the diagnosis of thymoma; in the other seven patients with thymoma, it developed an average of 7 years after thymectomy (range 1–22 years).

View this table:
Table 3

Associated immune‐related diseases

DiseaseGroup A (n = 42)Group B (n = 18)
n (%)n (%)
None9 (21)12 (67)
Neuropathy (see Table 5)11 (26)1 (6)
Myasthenia gravis + thymoma + AChR antibodies 6 (14)1 (6)
 Benign thymoma31
 Malignant thymoma30
Myasthenia gravis present3 (7)1 (6)
 AChR+20
 AChR11
Thymoma + AChR antibodies, myasthenia gravis absent2 (5)1 (6)
Diabetes mellitus4 (10)0
 Insulin20
 Non‐insulin20
Lung cancer4 (10)1 (6)
 Small cell30
 Adenocarcinoma11
Rheumatoid arthritis2 (5)0
Amyloidosis01 (6)
Systemic lupus erythematosus1 (2)0

Five patients developed lung cancer (Table 3), providing further evidence that motor nerve hyperexcitability syndromes can be paraneoplastic. Four patients were cigarette smokers, three had a small cell lung carcinoma and the other two had an adenocarcinoma. The neurological disorder pre‐dated the diagnosis of the tumour by an average of 2.3 years (range 1–4 years). None of the patients had any other neurological paraneoplastic syndrome or serum antibodies to intracellular neurone proteins such as Hu.

Penicillamine was previously linked to neuromyotonia in a patient with rheumatoid arthritis (Reeback et al., 1979). Two group A patients with rheumatoid disease had used penicillamine, but stopped the drug ∼4 years before muscle overactivity appeared, suggesting that the disorder was not penicillamine‐induced.

Twenty‐four per cent and 39% of patients in groups A and B, respectively, had one or more first‐degree relatives with an autoimmune disease—usually thyroid disease or diabetes.

Serum VGKC antibodies and other autoantibodies

We found raised titres (≥100 pmol/l) of serum VGKC antibodies in 15 out of 39 (38%) of group A patients and five out of 18 (28%) of group B patients tested (Fig. 2). Thymoma occurred in both groups, and the VGKC antibody titre was raised in eight of these 10 patients. Of those who had a raised titre without thymoma, there were single cases (all in group A) of myasthenia gravis, idiopathic axonal neuropathy and systemic lupus erythematosus.

Fig. 2 Scatter plot of serum VGKC antibody titres in 39 group A and 18 group B patients. Patients with thymoma are shown separately. Titres were considered positive if more than the mean plus 3 SD of control sera (≥100 pM).

In the 20 patients positive for VGKC antibodies, we found no correlation between the titre and the severity of the clinical or EMG features.

Other serum autoantibodies were detected in both groups and usually reflected the presence of an additional autoimmune disease (Table 4). The patients with raised thyroid or reticulin antibodies and the three patients with positive antinuclear antigen were otherwise well. The patient with thymoma, AChR and glutamic acid decarboxylase antibodies had no clinical evidence of myasthenia, stiff person syndrome or diabetes.

View this table:
Table 4

Other auto‐antibodies

AutoantibodyGroup A (n = 42)Group B (n = 18)
n (%)n (%)
No other autoantibody21 (50)15 (83)
AChR11 (26)2 (11)
Antinuclear antigen4 (10)0
Striated muscle3 (7)0
Thyroid3 (7)1 (6)
Rheumatoid factor2 (5) 0
Reticulin2 (5)0
Glutamic acid decarboxylase1 (2)0
Double‐stranded DNA1 (2)0

Nerve conduction study findings

The majority of patients in group A and all but one in group B had no electrophysiological evidence of either a demyelinating or axonal neuropathy (Table 5). Details of those with abnormal nerve conduction are given in Table 5; in most cases, the neuropathy was idiopathic, axonal and subclinical. Of the ten patients with abnormal nerve conduction who were tested for VGKC antibodies, only one patient with an idiopathic mild axonal neuropathy had a raised titre.

View this table:
Table 5

Associated peripheral neuropathy

NeuropathyGroup A (n = 42)Group B (n = 18)
n (%)n (%)
None31 (74)17 (94)
Idiopathic axonal6 (14)1 (6)
 Sensorimotor51
 Sensory10
Diabetes‐associated axonal2 (5)0
 Sensorimotor20
Chronic inflammatory demyelinating polyneuropathy (glycolipid antibody negative)2 (5)0
Systemic lupus erythematosus‐associated axonal1 (2)0

CSF analysis

CSF abnormalities were found in 11 out of 27 group A patients (41%). Six had oligoclonal bands, of whom two had an associated lung cancer. Five others had a mild to moderate increase in total protein (range 0.52–0.9 g/decilitre), including the two patients with chronic inflammatory demyelinating neuropathy. All patients had a normal CSF lymphocyte count. The only group B patient tested had a normal CSF.

EMG findings

Group A patients had, by definition, myokymic EMG discharges [i.e. doublet, triplet or multiplet motor unit (or partial motor unit) discharges], as well as fasciculations (single spontaneous motor unit discharges) and fibrillations (single spontaneous muscle fibre discharges). The occurrence of doublet discharges was the commonest abnormality; these had a mean intraburst frequency of 168 Hz (range 40–280 Hz). Denervation was only observed in patients who had abnormal nerve conduction studies.

Prolonged surface EMG recordings over 30 min from eight different limb muscles in six patients revealed that the greatest number of myokymic discharges occurred in distal rather than proximal muscles. Ten was the maximum number of different motor units (or partial motor units) showing myokymic activity detected on surface EMG recording. The extent of the electrophysiological abnormality did not relate to the clinical severity of muscle hyperactivity.

Twenty‐four group A patients (57%), but none from group B, had prolonged after‐discharges following median nerve stimulation at 5 Hz. This suggests that the peripheral nerves are more excitable in the former group. All group B patients had fasciculations; their EMG studies were otherwise normal and no spontaneous repetitive discharges were observed.

Psychiatric features

Eleven group A patients (26%) reported psychiatric symptoms—most commonly isolated mild personality change, insomnia or irritability. One patient also had visual hallucinations and one developed a delusional state. Five of them had raised serum VGKC antibody titres and the two patients with the highest serum titres in our series of 60 patients were the two group A patients with the most florid psychiatric symptoms. Eight of these 11 group A patients who underwent CSF analysis had a mildly raised total protein, but no oligoclonal bands were detected.

A similar proportion of group B patients (22%) reported mild psychiatric symptoms. One of these patients also had vivid dreams and only in this patient was the VGKC titre raised.

Other investigations

Creatine kinase was raised in about half the patients tested in both groups. All patients had normal serum calcium and phosphate, and those with pseudotetany also had normal magnesium. Although neuromyotonia can occur in hyperthyroidism (Harman and Richardson, 1954), all our patients were clinically and biochemically euthyroid.

Discussion

We divided our 60 patients with features of muscle hyperexcitability into two groups based on the presence (group A) or absence (group B) of myokymic EMG discharges (doublet, triplet or multiplet motor unit discharges). Group A comprised patients who would be clinically classified as having either undulating myokymia (Denny‐Brown and Foley, 1948) or neuromyotonia (persistent muscle contractions or pseudomyotonia). Group B patients broadly fitted the clinical definition of Cramp–Fasciculation Syndrome (Tahmoush et al., 1991). However, after‐discharges were less frequent in our patients, possibly because we only studied one median nerve from each patient, whereas Tahmoush et al. (1991) studied four peripheral nerves and found after‐discharges more frequently in the lower limbs.

The most striking clinical finding was the similarity between the two groups in the clinical, electrophysiological and immunological features. Indeed, it would be difficult to discriminate between them without the aid of EMG studies. We found no difference in the age of onset and sex ratio, and no qualitative difference in the presenting symptoms, signs and topography. The only consistent clinical distinction was quantitative: group A patients typically presented with a greater number of motor features that tended to be more severe and more widespread than in group B patients.

Both groups had strong associations with autoimmune diseases, and could associate with lung cancer. As in the case of other antibody‐mediated disorders such as the Lambert–Eaton myasthenic syndrome, these associations point to a likely immune‐mediated pathogenesis in both groups (O’Neill et al., 1988). The most striking immunological finding, however, was that many patients in both groups had raised serum VGKC antibody titres. Our previous studies of a selected group of patients with undulating myokymia or neuromyotonia, and EMG doublet or multiplet motor unit discharges, established that the cause of their nerve hyperexcitability was likely to be an antibody‐mediated potassium channelopathy (Sinha et al., 1991; Newsom‐Davis and Mills, 1993; Shillito et al., 1995; Hart et al., 1997). Others have confirmed these findings (Sonoda et al., 1996; Arimura et al., 1997; Nagado et al., 1999). Thus, the presence of VGKC antibodies in 28% of patients of group B implies that this Cramp–Fasciculation Syndrome like syndrome can also be an autoimmune channelopathy.

The distinction between those with myokymic EMG discharges (group A) and those without is not immutable. Two of our patients, who initially would have been classified as group B, had developed myokymic motor unit discharges on their second or third EMG study, indicating that the severity of the electrophysiological phenotype and thus the degree of nerve hyperexcitability may vary over time. In addition, a single EMG study is liable to spatial as well as temporal sampling error. Wider or longer muscle sampling in group B patients might have detected a repetitive discharge that would change their category to group A. Taken together with our clinical and immunological data, this suggests that the EMG distinction between these syndromes is primarily quantitative.

Although overall only 35% of patients had raised VGKC antibody titres detected by the 125I‐α‐dendrotoxin immunoprecipitation assay, this is likely to be an underestimate. We have previously discussed the relative insensitivity of this assay compared with a molecular‐immunohistochemical assay that detects serum binding to frozen sections of Xenopus oocytes injected beforehand with cRNA for an individual VGKC α subunit (Hart et al., 1997). This molecular assay has several advantages: (i) a ligand is not required to label the VGKC α subunit; (ii) the antigen (a single α subunit isoform) is expressed at high concentration in the oocyte cytoplasm; and (iii) antibody binding can be studied without the need to extract and denature the antigen with detergent. In previous studies of patients with acquired neuromyotonia or undulating myokymia, 26 out of 30 patients (87%) tested positive at serum dilutions of 1 : 128 to 1 : 2500 (reviewed by Hart, 2001). Although it was not possible to test all of the present patients’ sera on this molecular assay, 17 out of 19 group A patients (89%) tested in this assay bound to at least one of three expressed human VGKC α subunits (KCNA1, KCNA2 or KCNA6; data not shown). Moreover, seven of these positive sera were from patients who tested negative by the immunoprecipitation assay. We speculate, therefore, that the serum of most of our patients may have contained VGKC antibodies, and that these antibodies contributed to the pathogenesis of the nerve hyperexcitability. The 125I‐α‐dendrotoxin immunoprecipitation assay, however, is still the only assay for VGKC antibodies that is routinely available and provides a useful screening test for the diagnosis of autoimmune PNH.

Our EMG data suggest that, in many patients, PNH was more evident distally than proximally. This is in keeping with the previous evidence from detailed peripheral nerve excitability studies of 20 of the patients reported here that spontaneous activity is often generated focally or multifocally at sites distant from the recording electrode over the trunk of the nerve (Maddison et al., 1999; Kiernan et al., 2001). Thus, our present findings together with the evidence provided by these excitatory techniques and previous histological and conventional electrophysiological findings (Isaacs, 1967; Deymeer et al., 1998; Newsom‐Davis and Mills, 1993; Vincent, 2000) suggest that, in most patients, the likely locus for the generation of spontaneous discharges is at the motor nerve terminal or intramuscular arborization. This was proposed originally by Isaacs (1961). At these sites, there are VGKCs producing fast K+ currents unprotected by either the blood/nerve barrier or myelin sheath, and thus potentially more vulnerable to antibody‐mediated autoimmune attack.

The majority of our patients had no clinical or electrophysiological evidence of an axonal or demyelinating neuropathy in which PNH can occur, either at presentation or during follow‐up lasting 1–10 years. This finding emphasizes that acquired nerve hyperexcitability is not a single disease process, but a response to peripheral nerve dysfunction or damage arising from several different causes. The term axonal neuropathy was proposed to distinguish the majority of patients who have isolated nerve hyperexcitability from those who have an associated peripheral neuropathy (Gutmann and Gutmann, 1996). We found VGKC antibodies at presentation in one patient with idiopathic axonal peripheral neuropathy, demonstrating that these antibodies can co‐exist with axonal neuropathy and suggesting that they may contribute to hyperexcitability even when there is nerve damage from other causes. Although two patients who presented with isolated nerve hyperexcitability later developed a neuropathy confirmed by nerve conduction studies during follow‐up, both had diabetes. Thus, our patients provide no evidence that chronic nerve hyperexcitability caused by VGKC antibodies directly leads to a neuropathy alone.

Our analysis of the disorders associating with PNH raises other issues. About 25% of our patients had CNS symptoms. Although many of the milder problems such as anxiety and insomnia may have been coincidental or secondary to the morbidity caused by the symptoms of PNH, two patients had severe psychiatric features requiring anti‐psychotic drug treatment. This combination of PNH and severe psychiatric problems including delusions and visual hallucinations is called Morvan’s syndrome (Serratrice and Azulay, 1994; Barber et al., 2000). These two patients had the highest serum VGKC antibody titres of the entire series. However, neither had CSF oligoclonal bands and VGKC antibodies have only infrequently been detected in CSF samples from other patients with Morvan’s syndrome (Liguori et al., 2001; A. Vincent, unpublished results). One possible interpretation of these findings is that, at least in our two patients, circulating rather than CNS VGKC antibodies contributed to the CNS manifestations of Morvan’s syndrome.

Issues of practical importance arise from the strong association between PNH and immune‐related diseases, such as myasthenia gravis with thymoma or lung cancer. First, our findings suggest that all patients presenting with acquired PNH should have a serum autoantibody screen that includes VGKC and AChR antibodies, plus glucose and thyroid function tests to help exclude other autoimmune diseases. Secondly, because PNH can be paraneoplastic, it is important to search for an underlying thymoma or lung cancer. Although nerve hyperexcitability presented in most of our thymoma patients after the tumour had been diagnosed, in the remainder the development of the neurological disorder might have enabled earlier identification of the neoplasm. Patients developing PNH over the age of 40 years who are seropositive for AChR and VGKC antibodies are most at risk. Eighty per cent of our thymoma patients had VGKC antibodies detected by the 125I‐α‐dendrotoxin immunoprecipitation assay, suggesting that the assay may be particularly sensitive in this subgroup of PNH patients. CT imaging of the chest should also be considered for patients who are cigarette smokers. In our patients, the lung tumours presented up to 4 years after the onset of nerve hyperexcitability. This long latency is also seen in the Lambert–Eaton myasthenic syndrome where small cell lung carcinoma can declare itself up to 4 years after the neurological disorder (O’Neill et al., 1988). Serial CT or chest MRI should be considered in high risk patients, i.e. cigarette smokers with late‐onset PNH.

Since the first electrophysiological characterization of generalized PNH by Denny‐Brown and Foley (1948), a new term for the various clinical manifestations has been proposed on average every 2.7 years. This plethora of eponymous and descriptive names can lead to nosological uncertainty (see Gutmann et al., 2001). Moreover, while these different terms are useful initially in defining the diverse clinical manifestations (e.g. undulating myokymia, neuromyotonia, Cramp–Fasciculation Syndrome), they have the limitation of not including the underlying pathogenesis. Thus, we endorse the use of the generic term ‘peripheral nerve hyperexcitability syndromes’ to describe these conditions as proposed by Tahmoush and colleagues (Tahmoush et al., 1991). But the results of the present study now allow us to add a new pathogenetic dimension to the classification based on the strong autoimmune associations found in each of these syndromes and, specifically, the detection of VGKC antibodies in many of the patients. Rather than being based on the clinical or EMG phenotype, the classification proposed in Table 6 takes into account our data and that of others on the clinical associations and likely pathogenesis (reviewed by Lance et al., 1979; Auger et al., 1984; Newsom‐Davis and Mills, 1993; Auger, 1994; Jamieson and Katirji, 1994; Hart et al., 1999; Vincent, 2000). The classification reinforces the point that PNH seen in the clinic or EMG laboratory can result from a variety of pathogenic processes, with important implications for management. Thus, acquired forms are classed as autoimmune, toxic or degenerative, and hereditary forms are defined by their underlying genetic disorder. This classification has the added advantage of being flexible. It can accommodate new information about the molecular pathogenesis of the syndromes without the need for a fundamental reclassification and should facilitate the testing of new therapies.

View this table:
Table 6

Clinicopathogenetic classification of acquired syndromes of generalized PNH

Autoantibody‐mediated or autoimmune‐associated
Isolated*
Associated with CNS features; Morvan’s syndrome*
Paraneoplastic
 Thymoma with or without myasthenia gravis*
 Small cell lung carcinoma with or without neuropathy*
 Adenocarcinoma*
 Hodgkin’s lymphoma,
 Plasmacytoma with IgM paraproteinaemia
Associated with idiopathic peripheral neuropathy*
Associated with other autoimmune disorders
 Myasthenia gravis without thymoma*
 Diabetes mellitus: insulin or non‐insulin dependent, with or without diabetic neuropathy
 Chronic inflammatory demyelinating neuropathy
 Guillain–Barré syndrome*
 Addison’s disease with demyelinating neuropathy
 Rheumatoid disease
 Systemic lupus erythematosus*
 Systemic sclerosis*
 Hyperthyroidism, hypothyroidism*
 Coeliac disease
 Amyloidosis with or without paraproteinaemia
 Penicillamine‐induced in patients with rheumatoid disease
Non‐immune mediated
Toxins: herbicide, insecticide, toluene, alcohol, timber rattle snake envenomation
Drugs: gold
Idiopathic peripheral neuropathy
Anterior horn cell degeneration as part of motor neurone disease
Identified gene mutations
Voltage‐gated potassium channel (KCNA1)
 Familial episodic ataxia type 1
 Sporadic
Peripheral myelin protein (PMP 22)
 Hereditary neuropathy with pressure palsies
 Hereditary motor sensory neuropathy type 1a
Other hereditary diseases
Spinal muscular atrophy
Schwartz–Jampel syndrome
Heredofamilial without overt neuropathy
Other hereditary neuropathies—motor or sensory–motor

*Disorders in which VGKC antibodies have been detected.

Acknowledgements

We thank Dr Brian Tedman and his colleagues in the Department of Neurophysiology at The Walton Centre for their help with the investigation of the Liverpool patients.

References

View Abstract