Bilateral foot drop, paraesthesiae, and absent tendon reflexes in the lower limbs are, for the clinical neurologist, the hallmarks of a duration dependent sensory-motor neuropathy. We report a patient in which this clinical picture was sustained by the combination of an atypical distal presentation of myasthenia gravis with a probable immunomediated sensory neuronopathy.

A 69 year old woman presented with a three month history of progressive walking difficulties and paraesthesiae in the lower limbs. Examination showed bilateral foot drop with pronounced weakness of the tibioperoneal muscles (MRC= 2) and posterior leg muscles (MRC= 3), slight weakness in arm abduction (MRC=4+), tactile and pain distal sensory loss, and absent tendon jerks in the lower limbs. There were no oculobulbar symptoms and signs nor ataxia.

Motor conduction velocities and compound muscle action potential (CMAP) amplitudes were normal (ulnar= 57 m/s, 9.7 mV; peroneal= 44 m/s, 3.9 mV). Sensory conduction velocities were slowed with reduced amplitude sensory nerve potentials (ulnar=44 m/s, 5 μV; sural= 30.6 m/s, 1.8 μV). H reflexes were absent with normal latency F responses. Tibialis anterior muscle EMG did not show spontaneous activity and recruitment was full but during prolonged maximal voluntary activation there was a pronounced reduction of the interferential pattern amplitude. Quantitative motor unit analysis showed a 20% reduced mean duration with 4% of polyphasics. Repetitive stimulation at 3 Hz documented in the ulnar nerve-abductor digiti minimi system a 10% decrement whereas in the peroneal nerve-extensor digitorum brevis system the decrement was 33%. Amplitude CMAPs in the abductor digiti minimi after 15 seconds of maximal voluntary activation increased by 29%. A tensilon test, assessing strength in the tibioperoneal muscles, was strongly positive. Sural nerve biopsy showed a loss of myelinated fibres. Anti-AChR antibody titre was 3.3 nM (normal value<0.8). Voltage operated calcium channel antibodies were negative. Antibodies (IgG and IgM) against gangliosides (GM1, GD1a, GD1b, GM2) , SGPG, and sulphatides were negative. Chest CT excluded a thymoma or a lung cancer. A screening for breast, ovarian, and other gynaecological malignancies was negative. Tumorous markers were negative. Foot drop was greatly improved with pyridostigmine bromide and prednisone (75 and 25 mg every other day) and after three months the patient did not complain of paraesthesiae and could walk on her toes and stand on her heels.

By indirect immunofluorescence,1 the patient’s IgG strongly reacted with the cytoplasm of rat small size dorsal root ganglia neurons up to a dilution of 1:1280, whereas it did not react to large size dorsal root ganglia neurons (figure). With an indirect immunoperoxidase technique,1 using sections of rat cerebellum, the patient’s serum strongly immunostained the cytoplasm of Purkinje cells in a fine granular pattern up to a dilution of 1:2560. The patient’ serum did not react with neurons of the molecular layer and Golgi neurons. To determine whether the binding to Purkinje cells and dorsal root ganglia neurons was due to the same antibody, the patient’s serum was adsorbed either with whole cerebellum homogenate or with dorsal root ganglia homogenate. The pretreatment with each of the two homogenates removed the tissue binding activity, whereas liver homogenate was without effect. In the western immunoblot of whole human cerebellum homogenate, the IgG from our patient reacted with a 50 kDa protein band at dilutions up to 1:8000 (figure). Using whole dorsal root ganglia homogenate, the IgG from our patient reacted with a 50 kDa protein band at dilutions up to 1:1000. The pretreatment of serum either with whole cerebellum homogenate or with whole dorsal root ganglia homogenate removed their blot binding activity, whereas liver homogenate was without effect.

• Download figure
• Open in new tab
• Download powerpoint

Top: binding of patient’s serum at a dilution of 1:1280 to an unfixed frozen section of rat dorsal root ganglia by immunofluorescence microscopy. The patient’s IgG intensely immunostains the cytoplasm of small size dorsal root ganglia neurons (arrows) but not with large size dorsal root ganglia neurons (arrowhead) (originally×170). Middle: control serum does not bind and stain the cytoplasm of rat dorsal root ganglia(originally×170). Lower: lane 1: immunoblot of human whole cerebellum homogenate shows that at a serum dilution of 1:8000 the patient’s IgG (lane 1) binds to a 50 kDa protein. Lane 2: positive control with anti-Hu antibodies binding to a 40 kDa protein. Lane 3: negative control. Lane A: whole cerebellum homogenate; lane B: molecular weight standard. Lanes 1–3, peroxidase conjugated assay. Lanes A and B, amido black staining.

Clinical presentation in this patient pointed to a sensory-motor neuropathy. Neurophysiological studies documented an axonal sensory neuropathy but normal motor conduction. Normal CMAP amplitudes and the borderline reduced mean duration value of motor unit potentials in tibialis anterior muscle indicated that the foot drop did not have a neurogenic origin. Repetitive stimulation showed an unequivocal decremental response in the peroneal nerve (whereas the ulnar nerve was still in the normal range) indicating a defect in neuromuscular transmission, more pronounced in leg muscles. The normal CMAP amplitudes, the absence of incremental response, and the negativity of voltage operated calcium channel antibodies ruled out a Lambert-Eaton myasthenic syndrome. Myasthenia gravis was confirmed by the raised titre of serum anti-AchR antibodies and the positive tensilon test. In myasthenia gravis distal limb muscles are reported to be almost never affected in absence of oculobulbar signs.2 Peroneal nerve repetitive stimulation, as recently described by Oh et al,3 was crucial to the confirmation of myasthenia gravis in this patient.

Clinical, electrophysiological, and histological findings also showed an axonal sensory neuropathy. Although there are no electrophysiological or morphological ways to distinguish between an axonal sensory neuropathy and a sensory neuronopathy, the reactivity of our patient’s IgG against small size dorsal root ganglia (figure) suggests a primary, probably immunomediated, involvement of sensory neurons. The patient’s serum strongly immunostained, in a fine granular pattern, the cytoplasm of Purkinje cells and in western immunoblotting the IgG reacted with a 50 kDa protein band present in whole cerebellum and dorsal root ganglia homogenates (figure). These immunostaining and immunoblot findings are clearly different from the features required for the identification of anti-Yo and anti-Hu antineuronal antibodies.4 5

In conclusion this case is unique for several reasons:

(1)

Clinical presentation pointed to a sensory-motor neuropathy but the foot drop was sustained by an atypical, distal presentation of myasthenia gravis.

(2)

Peroneal nerve repetitive stimulation (not usually performed in EMG laboratories) played an important part in confirming myasthenia gravis.

(3)

Myasthenia gravis has been reported to be associated with other autoimmune disorders but never to a sensory neuronopathy with antibodies reacting with dorsal root ganglia neurons and Purkinje cells.

(4)

50 kDa antineuronal antibodies reacting with small size dorsal root ganglia neurons and in a fine granular pattern with the cytoplasm of Purkinje cells have never been described previously.