Focal upper limb demyelinating neuropathy

Transcription

1 Focal upper limb demyelinating neuropathy P. K. Thomas, 1 D. Claus, 4 A. Jaspert, 4 J. M. Workman, 1 R. H. M. King, 1 A. J. Lamer, 2 Milne Anderson, 2 J. A. Emerson 3 and I. T. Ferguson 3 Brain (1996), 119, Department of Clinical Neurosciences, Royal Free Hospital School of Medicine, London, the ^Midland Centre for Neurology and Neurosurgery, ^Southmead Hospital, Bristol, UK and the A Neurologische Klinik der Universita't Erlangen-Nurnberg, Germany Correspondence to: Professor P. K. Thomas, Clinical Neurosciences, Royal Free Hospital School of Medicine, Rowland Hill Street, London NW3 2PF, UK Summary Observations are presented on nine selected patients with chronic upper limb demyelinating neuropathy to illustrate the range of manifestations that may be observed. In three, the involvement was purely motor, in five, mixed motor and sensory and, in one, virtually purely sensory; in seven the symptoms were unilateral and in two bilateral. The presence of reduced nerve conduction velocity and conduction block and the response to treatment in seven of the cases indicate that they represented examples of chronic inflammatory demyelinating polyneuropathy (CIDP) with focal involvement. This was confirmed by nerve biopsy in two cases. The presentation in one patient was accompanied by forearm swelling initially suspected of being a tumour but shown to be due to muscle hypertrophy. This was probably the consequence of recurrent muscle cramps and fasciculation and possibly neuromyotonia. The patient with predominant sensory involvement restricted to the upper limbs demonstrates that sensory CIDP can present focally. In one patient with monomelic motor and sensory involvement, nerve biopsy showed multifocal areas of hypertrophic demyelinating neuropathy distally in the ulnar nerve without inflammatory infiltration. This patient failed to respond to therapy. Response in the others was satisfactory, although one patient with a monomelic motor neuropathy showed a severe deterioration after being given corticosteroids; he subsequently improved with intravenous human immunoglobulin therapy. Keywords: neuropathy; inflammatory demyelination Abbreviations: CIDP = chronic inflammatory demyelinating polyneuropathy; IVIg = intravenous immunoglobulin Introduction Chronic inflammatory demyelinating polyneuropathy was first recognized by Austin (1958), Hinman and Magee (1967), Thomas et al. (1969) and Dyck et al. (1975) as a diffuse condition that pursues either a relapsing or a chronic progressive course. Pathologically it is characterized by mononuclear inflammatory cell infiltrates, macrophagemediated demyelination and often hypertrophic Schwann cell proliferation (see Hughes, 1995). The condition of multifocal motor neuropathy with conduction block has been separated as a distinct condition (Lewis et al., 1982) but it is probably a variant of CIDP as intermediate cases are not infrequent. Sensory nerve biopsy from cases of multifocal motor neuropathy has shown demyelination (Corse et al., 1996). The focal occurrence of what appears to be a similar process to that of CIDP (Cusimano et al. 1988) was initially recorded for patients with hypertrophic brachial plexus neuropathy by Adams et al. (1965) but isolated limb nerves Oxford University Press 1996 may be affected. The patient reported by Mitsumoto et al. (1990) had a localized inflammatory demyelinating neuropathy affecting one tibial nerve. It is our experience that these cases can be confusing diagnostically. We therefore report a series of nine patients with focal upper limb demyelinating neuropathy selected from others whom we have seen, who demonstrate that the presenting features may be quite variable, as may their response to treatment. A short account of some of the findings has appeared in abstract form (Thomas et al., 1995). Methods The nerve conduction studies were performed by standardized techniques employing the control values documented by Kimura (1989) and Claus (1990). Conduction block was accepted using the criteria proposed by Cornblath et al. (1991) and Jaspert et al. (1995).

2 766 P. K. Thomas et al. Table 1 Nerve conduction results Case I Case 2 Case 4 Motor R median R ulnar R peroneal L median Motor (R) Median nerve Ulnar nerve Motor (R) Median Ulnar Peroneal Tibial Velocity/ Amplitude Velocity/ Amplitude latency latency Elbow/wrist (ms ') (or knee/ankle) Distal latency (ms) F wave latency (ms) CMAPs (mv) Proximal stimulation (elbow or knee) Distal stimulation (wrist or ankle) C7-0 0 Axilla/elbow 60 ms" mv 50 ms 2.5 mv Elbow/wrist 54 ms" mv 51 ms" mv Wrist/muscle 2.8 ms 2.4 mv 3.2 ms 3.5 mv F waves 0-0 MNCV (ms" 1 ) DML (ms) F latency (ms) CMAPs (mv) Distal stimulation Proximal stimulation II Sensory Median Ulnar Radial Sural Sensory (R) Median Ulnar Radial Sensory Right Left R L R L R L R L Median Ulnar Median Ulnar Amplitude (uv) Amplitude (uv) Amplitude ((tv) Latency to peak (ms) Latency/velocity 2.1 ms - 51 ms" Latency (ms) Serum anti-gmi ganglioside antibodies were estimated by ELISA (enzyme-linked immunosorbent assay). Except for Case 9 when this was undertaken in the University of Erlangen, they were kindly performed by Dr N. Gregson, Guy's Hospital, London, UK. An abnormal titre was accepted if the value exceeded the control mean value by 3 SDs. The nerve biopsy specimens were fixed in 3% glutaraldehyde in PIPES (piperazine-n-yv'-bis 2-ethane suphonic acid) buffer, postosmicated and, after dehydration, embedded in Durcupan. Semi-thin sections (0.5 J.m) were stained with thionin and acridine orange (Sievers, 1971) or toluidine blue. Ultra-thin sections were contrasted with methanolic uranyl acetate and lead citrate and examined in a Zeiss EM 902 electron microscope. Muscle strength was evaluated before and 10 days after treatment with intravenous immunoglobulin (IVIg) using the MRC (Medical Research Council) rating scale. Case reports Case 1 A man aged 52 years, previously in good health and with no family history of neurological disorder, developed progressive weakness of his right hand in The intrinsic muscles of the right hand and medial forearm became wasted. He developed numbness and tingling of his right fifth finger which later spread to all digits. He found that pressure above his right clavicle 'made him jump'. There were no symptoms referable to his other limbs or to his cranial nerves. On examination, the abnormal neurological findings were confined to the upper limbs. There was wasting and weakness of all the small hand muscles on the right. There was also slight wasting of the medial forearm muscles on the right with mild weakness of the long finger flexors and flexor pollicis longus. The tendon reflexes were absent in the right arm as was the left triceps jerk; the left biceps and brachioradialis jerks were depressed. The appreciation of light touch and pinprick was impaired over all the digits of the right hand and along the medial forearm. Joint position sense was reduced in the fingers on the right. The peripheral nerves were not thickened. The nerve conduction results are given in Table 1. They showed clear evidence of a proximal demyelinating neuropathy on the right with borderline abnormalities of sensory conduction on the left. MRI showed enlargement of the brachial plexus on the right. Biopsy of the plexus revealed a hypertrophic 'onion bulb' neuropathy (Fig. 1). No inflammatory infiltrates were detected. The patient was initially treated with prednisolone but continued to deteriorate. Azathioprine was added in He has since remained stable on this drug together with a maintenance dose of prednisolone. Case 2 A woman aged 26 years presented with a 4-year history of numbness in the right fifth finger which subsequently spread proximally in the arm, together with weakness of the right hand. Section of the tendon of scalenus anterior at another hospital was stated to have produced transient improvement. Following this the weakness in her arm and the sensory loss increased and similar but milder weakness developed in the left upper limb. On examination, there was diffuse wasting and weakness in the right upper limb involving proximal and distal muscles and milder predominantly distal weakness in the left upper limb. Apart from the left biceps and

3 Focal demyelinating neuropathy 767 Table 1 Continued Case 7 Case 6 Motor (R) Axilla/elbow Elbow/wrist Wrist/muscle Median nerve Ulnar nerve Velocity/latency Amplitude Velocity/latency Amplitude 58 ms~' 51 ms" ms 53 ms" mv 7.0 mv 7.0 mv 19 ms" ms Motor 0.5 mv 2.1 mv 7.2 mv F waves Right median nerve Left median nerve Velocity/latency Amplitude Velocity/latency Amplitude Elbow/wrist Wrist/muscle F waves 24 ms" ms 59 ms 0.2 mv 0.2 mv 22 ms" 1 14 ms 0.5 mv 0.7 mv 0 Motor Right ulnar nerve Elbow/wrist Wrist/muscle Sensory (R) Median Ulnar Left ulnar nerve Velocity/latency Amplitude Velocity/latency Amplitude 29 ms ' 7.2 ms 0.4 mv 1.2 mv 33 ms" ms 0.4 mv 0.4 mv Radial F waves Amplitude (nv) 15 Latency/velocity Sensory 56 ms"1 3.9 ms Amplitude Velocity 59 ms Median Ulnar Left radial Right Left Right Left mv 18 ms" 1 0 R = right; L = left; CMAP = compound muscle action potential: MNCV = motor nerve conduction velocity; DML = distal motor latency. v^ % J ^ ^ % TS^ * wv-"^ v \» * * V *' : * ' brachioradialis jerks, the upper limb tendon reflexes were lost. There was cutaneous sensory impairment affecting C3 to T2 dermatomes on the right and C5 to Tl on the left. There were no abnormal findings over the trunk or in the legs. Although there was little volitional contraction of the small hand muscles on the right, a vigorous mechanical response was obtained with electrical stimulation of the median nerves at the wrist and elbow. No response was obtainable with paraspinal stimulation over the nerve roots. The results for motor and sensory nerve conduction in the right upper limb are given in Table 1. Motor and sensory conduction in the lower limbs was entirely normal. Routine haematological and biochemical investigations were normal, including an autoantibody screen and serum protein electrophoresis. Testing for anti-gmi ganglioside antibodies was negative. MRI of the brachial plexus and cervical spinal cord was normal, as was computed tomographic myelography. The CSF was normal. The patient was treated with prednisolone, initially 50 mg orally on alternate days, together with 100 mg azathioprine daily. Steady improvement occurred. The prednisolone was withdrawn after 15 months, at which time the only residual signs were slight weakness of shoulder abduction on the right and absent tendon reflexes in the right arm.». Fig. 1 Case 1. Transverse section through biopsy specimen from right brachial plexus showing multiple 'onion bulb' whorls. The endoneurial spaces are considerably expanded. Epoxy resin section, toluidine blue staining. Bar = 25 (im. Case 3 A woman aged 64 years presented in April 1995 with a 10year history of slowly increasing weakness in her right hand so that she could no longer continue using her hand for

4 768 P. K. Thomas et al. A Fig. 2 Case 3. Electromyographic tracings with surface recording from right abductor digiti minimi on stimulation of ulnar nerve at the wrist (1), distal to the elbow (2), proximal to the elbow (3) and in the axilla (4), and also over Erb's point (5) and the C7 vertebra (6). The tracings in the left hand panel (A) were taken before a 5-day course of high dose human IVIg and show conduction block with dispersion of the compound muscle action potential between C7 and Erb's point which has recovered I day after the completion of the course (B). everyday activities. There were no associated sensory symptoms and no symptoms referable to her other limbs or cranial nerves. Her past and family history were noncontributory. Examination revealed some muscle fasciculation in the right hand but no wasting. There was severe weakness of all the small hand muscles on the right and lesser weakness for finger extension. The tendon reflexes were depressed in the right arm. Otherwise neurological examination was negative. There was no sensory loss. Routine laboratory investigations were negative including CSF examination and testing for anti-g M) ganglioside antibodies. Electromyography of the right first dorsal interosseous, triceps and deltoid muscles showed evidence of denervation. Normal findings were obtained in the left extensor digitorum communis and both tibialis anterior muscles. Motor nerve conduction velocity in the right median nerve in the forearm was borderline at 46 ms~'. It was normal in the right ulnar nerve at 66 ms~' but there was evidence of conduction block between the spinal roots (stimulation over C7 vertebra) and Erb's point (Fig. 2A). Motor conduction velocity in the right peroneal nerve was normal (49 ms~') as was sensory conduction in the right radial and sural nerves (velocities 67 and 48 ms~', amplitudes 9.5 and 16.8 ^V). A right ulnar sensory action potential recorded later (June 1995) was slightly reduced in amplitude (5.3 iv) but had a normal velocity (53 ms~'). The patient was treated with a 5-day course of IVIg (30 g day" 1 ) with rapid improvement in muscle strength and disappearance of conduction block (Fig. 2B). Because of deterioration at the end of May 1995, further IVIg was given (30 g on one day), again with rapid improvement so that her hand could be used for most everyday activities. The same sequence was repeated 4 weeks later. Case 4 A man aged 46 years presented with a 12-month history of progressive weakness of the right hand. On examination, the abnormalities were confined to the right upper limb. There was wasting of the thenar muscles and interossei, mild weakness of triceps and moderate weakness of all muscles below the elbow. The brachioradialis and triceps jerks were depressed. There was no sensory loss. Routine haematological and biochemical investigations, including an autoimmune screen and serum protein electrophoresis, were normal, as were a chest X-ray, MRI of the brachial plexus and cervical spinal cord and a cervical radiculogram. The CSF protein concentration was borderline (0.5 g I" 1 ). Nerve conduction studies demonstrated severe prolongation of median and ulnar nerve F wave latencies on the right, but otherwise no significant abnormalities {see Table 1). The patient commenced treatment with prednisolone, 60 mg orally on alternate days, together with 150 mg azathioprine daily. Within 3 weeks severe deterioration of the weakness occurred, the right hand becoming useless. Treatment was discontinued, following which there was slow recovery to the previous strength. The patient was then given a 5-day course of IVIg. Improvement of strength in the right hand began within 48 h and a satisfactory response was maintained for 3 months. Further courses have been given subsequently, again with a satisfactory response. Case 5 A woman aged 71 years developed tingling paraesthesiae affecting the lateral three digits of her right hand in Median nerve decompression was performed at the wrist without benefit. In 1990 she became aware of weakness for flexion of the right thumb, index and middle fingers. Nerve conduction studies suggested a median nerve lesion in the region of the elbow but surgical exploration was negative. The weakness in the hand continued to progress and she began to experience cramps in her forearm muscles. Painless 'locking' of her grip occurred when using her hand. Focal

5 Focal demyelinating neuropathy Fig. 3 Case 5. Photographs of forearms showing enlargement of the flexor muscles of the right forearm. swelling of the upper forearm on the ventromedial aspect was noticed. On examination the abnormal findings were found to be confined to the right upper limb. There was diffuse enlargement of the forearm flexor muscle mass (Fig. 3) with weakness of the long flexors innervated by the median nerve. Abductor pollicis brevis and opponens pollicis were wasted and weak. Muscles supplied by the ulnar and radial nerves were normal and there were no abnormalities of the tendon reflexes or sensation. Elbow-wrist motor nerve conduction velocity in the right median nerve was reduced at 35 ms~' and the terminal latency was increased (5.8 ms). The thenar compound muscle action potential was 2.9 and 2.5 mv with stimulation of the median nerve at the wrist and elbow, respectively. No F waves were detectable. The right index finger-wrist median sensory nerve action potential was 13 iv in amplitude with a latency of 3.1 ms; no potential was recordable at the elbow. Motor conduction was normal in the left median nerve as was sensory conduction in both ulnar nerves. Needle electromyography of the right flexor digitorum superficialis and abductor pollicis brevis showed reduced motor unit recruitment in both muscles but particularly in the latter and spontaneous fasciculation in the former. No myokymic discharges were recorded. MRI of the cervical spine, brachial plexus and upper arm was normal. Imaging of the forearm demonstrated that the swelling was related to muscular enlargement. Routine 769 Fig. 4 Case 6. Transverse section through biopsy specimen from dorsal cutaneous branch of right ulnar nerve showing focal areas of hypertrophic change in which there are multiple 'onion bulb' whorls. Epoxy resin section, thionin and acridine orange stain. Bar = 50 im. haematological and biochemical investigations were normal, as were serum protein electrophoresis and testing for antigm! ganglioside antibodies. The patient was treated by high dose IVIg following which her symptoms rapidly improved. She was then given prednisolone and has since remained asymptomatic on a small maintenance dose. The enlargement of the forearm muscles has subsided. Case 6 A girl aged 14 years, previously in good health and with no family history of neurological disorder, presented in 1993 with slowly progressive weakness and wasting of the right hand that had developed over the previous 2-3 years. Apart from occasional numbness in the right fifth finger there had been no sensory symptoms and she had had no symptoms in her other limbs. Examination showed normal cranial nerve function. There was wasting of the medial forearm flexor muscles and of all the small hand muscles, and weakness of triceps and all muscle groups below the right elbow, especially the small hand muscles. All the tendon reflexes in the right arm were absent but they were normal in the left arm. The knee jerks

6 770 P. K. Thomas et al. ' * r,7 \ r ~* *v ganglioside antibodies were not detected. Nerve conduction studies (Table 1) showed severely reduced motor conduction velocity in the ulnar nerve in the right upper arm with evidence of conduction block and loss of the ulnar sensory nerve action potential. A right dorsal ulnar sensory nerve biopsy was performed. Myelinated fibre density was within normal limits (8679 mm"2; normal range: mm"2) (Jacobs and Love, 1985), but some fibres showed abnormally thin myelin sheaths for axon diameter, suggesting remyelination or regeneration. No active demyelination was seen but there were a few fibres undergoing active degeneration of Wallerian type. Occasional regenerative clusters were present which had contributed to the preservation of fibre density. The most striking abnormality was the presence of multifocal areas of hypertrophic neuropathy (Fig. 4) with onion bulb formations composed of concentrically proliferated Schwann cells around axons. Confirmation that these were Schwann cells was demonstrated by positive immunocytochemical staining for S-100 protein (Fig. 5A), absence of epithelial membrane antigen staining (Fig. 5B) and by electron microscopy (Fig. 6). No inflammatory cells were observed in semi-thin resin-embedded sections or on electron microscopy and labelling for CD4, CD8 and B lymphocytes and for macrophages on frozen sections was negative. The blood vessels appeared normal, as did the perineurium and epineurium. Treatment with plasma exchange, IVIg and prednisolone in combination with azathioprine failed to produce any improvement. Case 7 v Fig. 5 Case 6. (A) Transverse section through biopsy specimen from dorsal cutaneous branch of right ulnar nerve reacted for S100 protein showing a positive reaction by the cells in the focal areas of hypertrophic neuropathy. (B) Same specimen reacted for epithelial membrane antigen showing a negative reaction by these cells but a positive reaction in the perineurium (pn). Frozen sections. Bar = 50 im. were depressed and the ankle jerks normal. There was no sensory loss apart from possible impairment of light touch on the right fifth finger. The peripheral nerves were not enlarged. Routine haematological and biochemical tests were normal, as was MRI of her cervical spine and neck. Anti-G,vu A man aged 64 years, previously in good health, developed tingling in the tips of all fingers of both hands in December This spread to his palms and was accompanied by a sensation of numbness that spread up to wrist level. He noticed no accompanying weakness and had no symptoms referable to his cranial nerves or to his lower limbs. He reported aching discomfort in his neck. Examination disclosed slight weakness for internal rotation at both shoulders, absent biceps and brachioradialis tendon reflexes and mild cutaneous sensory impairment in both hands. Routine haematological and biochemical tests were normal, including a serum vitamin B12 estimation. MRI of his cervical spine showed degenerative changes with encroachment on intervertebral foramina at midcervical level. He was supplied with a cervical collar without benefit. Because of intensification of his symptoms nerve conduction studies were performed in July The results for the upper limbs are given in Table 1. They showed evidence of demyelination in motor and sensory fibres. Lower limb (peroneal and tibial) and sensory (sural) nerve conduction was normal bilaterally. Examination then showed similar findings to those obtained earlier except that no weakness was detected and severe impairment of two-point

7 demyelinating neuropathy 771 ax Fig. 6 Case 6. Electron micrograph of transverse section through biopsy specimen from dorsal cutaneous branch of right ulnar nerve showing 'onion bulbs' (ob) composed of concentrically proliferated Schwann cells (Sc) surrounding demyelinated axons (ax). Bar = 1 )im. discrimination in his fingers bilaterally was noted. A left radial nerve biopsy was undertaken. Myelinated fibre density was reduced at 3930 mm"2. There was evidence of active demyelination on occasional fibres. Teased fibre studies (96 fibres) showed demyelination/remyelination in 39 (41%) and regeneration in four (4%). The remainder were normal. Collections of mononuclear inflammatory cells, mainly in relation to epineurial blood vessels were present. Immunocytochemical staining showed these to be CD 4 + and CD 8 + T lymphocytes. Labelling for B lymphocytes was negative. Testing for anti-gm ganglioside antibodies was negative. The patient was treated with a course of high dose IVIg with modest improvement, followed by prednisolone. By May 1995 his symptoms had largely resolved. He has continued on a maintenance dose of prednisolone, relapsing on attempts at withdrawal. Case 8 A man aged 18 years developed progressive weakness of both upper limbs in March 1990 followed by paraesthesiae of his finger tips after some months. EMG studies demonstrated chronic partial denervation in affected muscles. When examined in January 1991 he showed diffuse weakness without wasting in both upper limbs, more marked on the right, but preserved power in his legs. His tendon reflexes were normal and there was no sensory loss despite the presence of paraesthesiae in his finger tips. Needle EMG studies showed denervation in both upper limbs with borderline motor nerve conduction velocity in both ulnar nerves but mildly increased F wave latencies (34.0 and 34.6 ms on the right and left, respectively) and evidence of conduction block between Erb's point and the axilla. Investigations were otherwise normal including CSF examination and testing for serum anti-gvn ganglioside antibodies. The patient was treated with corticosteroids which was followed by rapid improvement in his symptoms. Because of the occurrence of side effects the dosage of prednisolone was reduced to 10 mg day"'. This led to the recurrence of upper limb weakness and the development of proximal weakness in the lower limbs. He was then given a course of IVIg with improvement within 2 days. He was maintained on repeated courses of IVIg at increasing intervals until October 1993 since when he has remained symptom-free without treatment.

8 772 P. K. Thomas et al. Case 9 A man aged 36 years presented with progressive weakness of the left arm for 7 years. He had no relevant past or family history of illness. Examination showed no muscle wasting. There was diffuse weakness in the left arm, maximal distally, but none in the other three limbs. The left brachioradialis and knee jerks and both ankle jerks were depressed. Sensation was intact. Needle electromyography of extensor digitorum communis, abductor pollicis brevis and abductor digiti minimi on the left showed fibrillation potentials, pseudomyotonic discharges and reduced motor unit recruitment. Motor nerve conduction velocity in the left ulnar nerve was 48 ms~' and there was focal conduction block between C7 paraspinal level and Erb's point. The latency to the forearm extensor muscles on stimulation of the radial nerve above the elbow was 11.0 ms on the left but 4.6 ms on the right. The CSF showed no pleocytosis but the protein content was 0.72 g ]~'. Anti-G M1 ganglioside antibodies were present in the serum at an elevated titre of 1:300. The patient was treated with high dose IVIg followed by substantial improvement of the weakness but with subsequent deterioration. He has since been treated by monthly repeat courses of IVIg, each with temporary benefit. Discussion The clinical features and biopsy findings in Case 1 are generally similar to those originally described by Adams et al. (1965), the brachial plexus biopsy specimen showing a typical hypertrophic neuropathy with concentric Schwann cell proliferation giving rise to 'onion bulbs'. Although not confirmed by nerve biopsy, Cases 2-5 and 9 are acceptable as examples of focal CIDP on the basis of their clinical and electrophysiological features and their response to treatment. In Case 7 the diagnosis was confirmed by nerve biopsy. The presentation with focal upper limb symptoms initially gave rise to diagnostic difficulty in five of them. In Case 1 a brachial plexus tumour was questioned at another hospital, and Case 2 was initially diagnosed, again at another centre, as having the thoracic outlet syndrome, treated by section of the scalenus anterior tendon. In Case 4 motor neuron disease was initially considered and in Case 7, cervical spondylosis. The evolution of the diagnosis in Case 5 is discussed later. Detailed nerve conduction studies, looking in particular for proximal slowing of conduction and conduction block, are crucial. Focal inflammatory demyelinating neuropathy can clearly remain localized for prolonged periods of time. In Cases 1 and 9, the duration so far extends over 9 and 10 years, respectively. Nevertheless it is likely that these cases form part of the spectrum of CIDP. Case 8 began with symptoms restricted to the upper limbs but these later involved the lower limbs and Case 9 showed lower limb reflex depression in combination with monomelic upper limb weakness. Why the brachial plexus or upper limbs should be affected in particular in some cases is uncertain. A negative result for anti-g M] ganglioside antibodies was obtained in six of the seven patients in this series who were tested. The presence of these antibodies therefore did not prove helpful in their investigation. Case 9 showed a weakly positive result. This patient had a monomelic demyelinating motor neuropathy with conduction block. The criteria for accepting the existence of conduction block were as recommended by Cornblath et al. (1991). The presence of antibodies to G M i gangliosides has been linked in particular to multifocal motor neuropathy with conduction block (Pestronk et al., 1988J. Although these antibodies have been considered possibly to play an aetiological role in the causation of the conduction block (Santoro et al., 1992; Uncini et al., 1993), recent experimental studies have failed to confirm this suggestion (Harvey et al., 1995). The clinical features in our Case 5 were unusual. The initial presentation was with median sensory loss confined to the hand which suggested the carpal tunnel syndrome, but the patient failed to respond to median nerve decompression. Later, when median-innervated forearm muscles became affected, median nerve entrapment at the elbow was suspected but surgical exploration was negative. The development of swelling in the forearm then led to the suspicion of a neural tumour. MRI indicated that the swelling was due to enlargement of the forearm muscles. Evidently this represented work hypertrophy secondary to muscle cramps and probable neuromyotonia. On one occasion the patient's fingers locked around the clothes line as she was hanging up the washing in the garden. She was unable to release her hand and had to call her husband to prise open her fingers. Ectopic impulses may arise in motor axons in focal inflammatory demyelinating neuropathy. The case in which the tibial nerve was focally involved, reported by Mitsumoto et al. (1990), displayed persistent myokymia of the calf muscles. The varied response to treatment is of interest and parallels that in diffuse CIDP. The long-term efficacy of treatment is difficult to establish as the natural history of these focal upper limb demyelinating neuropathies is not known. Cases 2, 7 and 8 responded satisfactorily to corticosteroids, relapsing on withdrawal in Cases 7 and 8. They were given in combination with azathioprine in Case 2 for its 'steroidsparing' effect. It is not possible to state from our experience in these cases whether the addition of azathioprine to corticosteroids is advantageous, although in Case 1 it is perhaps of interest that he continued to deteriorate on prednisolone alone but stabilized when azathioprine was added. In a small controlled trial, Dyck et al. (1985) failed to demonstrate benefit with azathioprine at a dosage of 2 mg kg" 1 day"' in combination with prednisone as compared with prednisone alone. Higher doses of azathioprine may be required (Hughes, 1995). The rapid deterioration in Case 4 on corticosteroid treatment with improvement on withdrawal is noteworthy. This has been described for generalized

9 Focal demyelinating neuropathy 113 demyelinating motor neuropathy (Donaghy et al., 1994). Our Case 4 is the first description known to us of its occurrence in focal inflammatory neuropathy, although in three of the four cases described by Donaghy et al. (1994) the involvement was asymmetric. The explanation for this phenomenon is uncertain, but it is perhaps analogous to the deterioration that may occur in patients with myasthenia gravis initially treated with high dose corticosteroids or adrenocorticotrophic hormone. It is therefore prudent to commence corticosteroid therapy in patients with demyelinating motor neuropathy at a low dose, followed by a gradual increase, or to use other forms of treatment. Cases 3, 4, 5, 8 and 9 responded very satisfactorily to IVIg. Only a modest benefit was observed in Case 7. It is clear that the presence of long-standing weakness does not preclude a good response to treatment. Case 3, in whom weakness had progressively increased over the course of 10 years, responded well to IVIg. Evidently this was related to the persistence of proximal conduction block with limited axonal degeneration, as originally described by Lewis et al. (1982). It is of interest that testing for anti-gmi ganglioside antibodies in this patient was negative. Case 6 showed weakness confined to the right upper limb associated with a patchy reduction of nerve conduction velocity into the demyelinating range accompanied by evidence of conduction block. The nerve biopsy findings were unusual in showing multifocal areas of hypertrophic neuropathy of onion bulb type. The absence of inflammatory cell infiltration and the failure to respond to treatment with corticosteroids, IVIg or plasma exchange raises the suspicion that she was not an example of focal inflammatory demyelinating neuropathy. The possibility of mosaicism from somatic mutation in a gene, a germ line mutation of which gives rise to a demyelinating neuropathy such as occurs in the genes for peripheral myelin protein 22 or P o protein {see Snipes and Suter, 1995), has to be considered. This seems unlikely in view of the presence of definite conduction block in this patient. Conduction block is not a feature of generalized inherited demyelinating neuropathies (Lewis and Sumner, 1982), but whether this argument can be extended to possible instances of mosaicism is uncertain. Case 6 is not an example of focal hypertrophic neurofibrosis (Simpson and Fowler, 1966; Bilbao et al., 1984) in which localized peripheral nerve lesions occur that are characterized by appearances superficially resembling hypertrophic neuropathy, but in which the onion bulb-like structures are composed of perineurial rather than Schwann cells. This was excluded in our case. In Case 7, the clinical manifestations were almost entirely sensory. Patients with generalized CIDP can at times show predominant or purely sensory involvement (Oh et al., 1992), although motor nerve conduction velocity is usually markedly reduced even in the absence of weakness. This patient shows that restricted upper limb demyelinating sensory neuropathy may also occur, although motor nerve conduction velocity was again reduced into the demyelinating range in the absence of significant weakness. Acknowledgements We wish to thank Dr Richard Wise for referring Case 1, Dr J. Gibbs for Case 5, Dr R. Weiser for Case 6 and Dr G. M. Stern for Case 7. We also wish to thank Professor F. Scaravilli for the brachial plexus biopsy in Case 1. The electron microscope was provided by the Muscular Dystrophy Group of Great Britain. Financial support to P.K.T. from the Leverhulme and Wellcome Trusts is gratefully acknowledged. References Adams RD, Asbury AK, Michelsen JJ. Multifocal pseudohypertrophic neuropathy. Trans Am Neurol Assoc 1965; 90: Austin JH. Recurrent polyneuropathies and their corticosteroid treatment: with five-year observations of a placebo-controlled case treated with corticotrophin, cortisone, and prednisone. Brain 1958; 81: Bilbao JM, Khoury NJS, Hudson AR, Briggs SJ. Perineurioma (localized hypertrophic neuropathy). Arch Pathol Lab Med 1984; 108: Claus D. Central motor conduction: method and normal results. Muscle Nerve 1990; 13: Cornblath DR, Sumner AJ, Daube J. Gilliatt RW, Brown WF, Parry GJ, et al. Conduction block in clinical practice. Muscle Nerve 1991; 14: Corse AM, Chaudry V, Crawford TO, Cornblath DR, Kund RW. Sensory nerve pathology in multifocal motor neuropathy. Ann Neurol 1996; Cusimano MD, Bilbao JM, Cohen SM. Hypertrophic brachial plexus neuritis: a pathological study of two cases. Ann Neurol 1988; 24: Donaghy M, Mills KR, Boniface SJ, Simmons J, Wright I, Gregson N, et al. Pure motor demyelinating neuropathy: deterioration after steroid treatment and improvement with intravenous immunoglobulin. J Neurol Neurosurg Psychiatry 1994: 57: Dyck PJ, Lais AC, Ohta M, Bastron JA, Okazaki H, Groover RV. Chronic inflammatory polyradiculoneuropathy. Mayo Clin Proc 1975; 50: Dyck PJ, CTBrien P, Swanson C, Low P, Daube J. Combined azathioprine and prednisone in chronic inflammatory-demyelinating polyneuropathy. Neurology 1985; 35: Harvey GK, Toyka KV, Zielasek J, Kiefer R, Simonis C, Hartung H-P. Failure of anti-g M! IgG or IgM to induce conduction block following intraneural transfer. Muscle Nerve 1995; 18: Hinman RC, Magee KR. Guillain-Barre syndrome with slow progressive onset and persistent elevation of spinal fluid protein. Ann Intern Med 1967; 67: Hughes RAC. Guillain-Barre syndrome and chronic inflammatory demyelinating polyradiculoneuropathy. In: Asbury AK, Thomas

10 774 P. K. Thomas et al. PK, editors: Peripheral nerve disorders 2. Oxford: Butterworth- Heinemann, 1995: Jacobs JM, Love S. Qualitative and quantitative morphology of human sural nerve at different ages. Brain 1985; 108: Jaspert A, Claus D, Grehl H, Kerling F, Neundorfer B. Wertigkeit der proximalen Leitungsblockuntersuchung in der Diagnostik entzundlicher Neuropathien. Nervenarzt 1995; 66: Kimura J. Electrodiagnosis in diseases of nerve and muscle: principles and practice. 2nd ed. Philadelphia: F. A. Davis, Lewis RA, Sumner AJ. The electrodiagnostic distinction between chronic familial and acquired demyelinative neuropathies. Neurology 1982; 32: Lewis RA, Sumner AJ, Brown MJ, Asbury AK. Multifocal demyelinating neuropathy with persistent conduction block. Neurology 1982; 32: Mitsumoto H, Levin KH, Wilbourn AJ, Chou SM. Hypertrophic mononeuritis clinically presenting with painful legs and moving toes. Muscle Nerve 1990; 13: Oh SJ, Joy JL, Kuruoglu R. 'Chronic sensory demyelinating neuropathy': chronic inflammatory demyelinating polyneuropathy presenting as a pure sensory neuropathy. J Neurol Neurosurg Psychiatry 1992; 55: Pestronk A, Cornblath DR, Ilyas AA, Baba H, Quarles RH, Griffin JW, et al. A treatable multifocal motor neuropathy with antibodies to GM1 ganglioside. Ann Neurol 1988; 24: Santoro M, Uncini A, Corbo M, Staugaitis SM, Thomas FP, Hays AP, et al. Experimental conduction block induced by serum from a patient with anti-gml antibodies. Ann Neurol 1992; 31: Sievers J. Basic two-dye stains for epoxy-embedded u.m sections. Stain Technol 1971; 46: Simpson DA, Fowler M. Two cases of localized hypertrophic neurofibrosis. J Neurol Neurosurg Psychiatry 1966; 29: Snipes GJ, Suter U. Molecular anatomy and genetics of myelin proteins in the peripheral nervous system. J Anat 1995; 186: Thomas PK, Lascelles R, Hallpike JF, Hewer RL. Recurrent and chronic relapsing Guillain-Barre polyneuritis. Brain 1969; 92: Thomas PK, Claus D Workman JM, King RHM. Focal upper limb inflammatory demyelinating neuropathy: varied clinical presentations and varied responses to treatment [abstract]. J Neurol 1995; 242 Suppl 2; S21. Uncini A, Santoro M, Corbo M, Lugaresi A, Latov N. Conduction abnormalities induced by sera of patients with multifocal motor neuropathy and anti-gml antibodies. Muscle Nerve 1993; 16: Received September 20, Revised January 12, Accepted January 19, 1996