Polyneuropathy. Facts and fancies.

Some conclusions are drawn from findings in 167 consecutive patients with the ordinary "garden variety" of polyneuropathy; the aetiology was unknown in 15%. Histological findings in sural nerves were related to clinical and electrophysiological abnormalities. In some patients with discrete clinical abnormalities, sensory and motor conduction and amplitudes of evoked sensory and muscle action potentials were normal, whereas the nerve biopsy showed slight but definite abnormalities. The reverse, abnormal nerve conduction and normal histological findings, did not occur. Histological findings were rarely, and electrophysiological findings were not, specific for the aetiology or type of a neuropathy. Thus, neither conduction studies nor conventional or single fibre electromyography can identify the underlying pathology: loss of large myelinated fibres (greater than 7 micrometers) was equally prominent in nerves with de- and re-myelination as in those without them. Paranodal and segmental demyelination in less than 20% of the teased fibres occurred as often in nerves with as in those without disproportionate slowing in conduction. When the recorded conduction velocity was equal to that to be expected from the fibres with the largest diameter, slowing in conduction could be explained by axonal degeneration ("proportionate" slowing, 79% of the nerves). When the recorded velocity was disproportionately slower than that expected from fibre diameter (21% of the nerves), causes other than loss of the largest fibres must be assumed to explain the slowing in conduction. Myelin abnormalities in more than 50% of the teased fibres were found only in nerves from patients with the hypertrophic type of peroneal muscular atrophy and in postgastrectomy neuropathy and can probably explain the marked disproportionate slowing in conduction. The material contained, however, only one patient with acute idiopathic polyradiculoneuropahy. In diabetic neuropathy, segmental demyelination was present in only 8 of 502 teased fibres (9 nerves), remyelination was present in 135 fibres, and could not explain the disproportionate slowing in conduction. The mechanism of disproportionate slowing, when it is not due to demyelination, is still obscure.     

Peripheral nerve structure and function in experimental diabetes

Observations have been made on the peripheral nerves of rats in which diabetes had been induced by streptozotocin or alloxan. Motor nerve conduction velocity was found to become reduced, the reduction developing within a few days in severely diabetic animals. Conduction velocity remained diminished during survival times of up to 1 year.Histometric studies of the myelinated fibre population of the sural and tibial nerves showed no loss of fibres or reduction in their calibre. Evidence of segmental demyelination was not detected and myelin/axon ratios were normal. Cation binding at the nodes of Ranvier was unaltered. No explanation for the reduced nerve conduction velocity in morphological terms was therefore obtained.Observations on the abdominal vagus nerve revealed no loss of unmyelinated axons or reduction in their calibre. Measurements of Schwann cell and endoneurial capillary basal laminal thickness showed no differences between diabetic and control animals. Ultrastructural examination of other features of the peripheral nerves failed to define any pathological alterations in the diabetic animals. It is therefore concluded that the extrapolation of biochemical findings in experimental diabetes in an attempt to explain the origin of human diabetic neuropathy, where there are associated structural changes, should be made with caution.     

Sensory action potentials and biopsy of the sural nerve in neuropathy.

In 167 consecutive patients with various types of neuropathy, the amplitude of the sensory potential and the maximum conduction velocity along the sural nerve were compared with conduction in other sensory nerves, and were related to structural changes revealed by nerve biopsy. Electrophysiological findings in the sural nerve were similar to those in the superficial peroneal and the median nerve, though the distal segment of the median nerve was normal in 20 per cent of the patients when it was abnormal in the sural nerve. Quantitation of histological findings was a more sensitive method than the electrophysiological study in that two-thirds of 33 patients with normal electrophysiology in the sural nerve showed mild loss of fibres or signs of remyelination in teased fibres. The amplitude of the sensory potential was grossly related to the number of large myelinated fibres (more than 7 micrometer in diameter). Considering the 95 nerves from which teased fibres were obtained, maximum conduction velocity was abnormal in half. In 18 of these nerves, slowing in conduction was due to axonal degeneration: the velocity was as to be expected from the diameter of the largest fibres in the biopsy ("proportionate slowing"). In 9 nerves slowing was severe and more marked than to be expected from loss of the largest fibres ("disproportionate slowing"); these nerves showed paranodal or segmental demyelination in more than 30 per cent of the fibres. In 16 nerves from patients with neuropathy of different aetiology neither loss of fibres nor demyelination could explain the moderate slowing. The cause of slowing in these nerves is unknown; other conditions are referred to in which slowing in conduction cannot be attributed to morphological changes. Finally, electrophysiological and histological findings are reported in some patients with neuropathy associated with malignant neoplasm, with rheumatoid arthritis, with polyarteritis nodosa, with acute intermittent porphyria and with cirrhosis of the liver.     

A micro-electrode study of peripheral neuropathy in man. Part 2. Responses to conditioning stimuli.

Surface, needle and micro-electrode recordings were obtained from sensory nerves of patients with various types of peripheral neuropathy. Changes in amplitude and conduction velocity of nerve action potentials were measured after a single conditioning stimulus and after tetanic stimulation for 2 min. In patients with hereditary forms of axonal degeneration (AD), recovery processes of nerve fibres of all conduction velocities were normal; in acquired forms of AD fibres with conduction velocity less than 30 m/sec had greater and more prolonged post-tetanic depression than control nerves of similar conduction velocity. Where neuropathy was associated with segmental demyelination (SD), fibres of all conduction velocities had prolonged recovery processes after both single and tetanic stimulation. The changes were especially marked at higher skin temperature, and were greater than the changes seen in nerves with acquired forms of AD. Finally, 2 sural nerves were studied during the process of Wallerian degeneration after a biopsy had been obtained proximally, and recovery processes did not change during the period of degeneration. Perceptual abnormalities were similar in AD and SD. It is suggested that changes in recovery processes of nerve fibres with segmental demyelination or regeneration after injury contribute to the perceptual abnormalities which occur in clinically encountered peripheral neuropathies.     

Improvement of endoneurial lipid abnormalities in experimental diabetic neuropathy by oxygen modification

Endoneurial hypoxia and a high frequency of closed capillaries have been found in chronic experimental diabetes and human diabetic sural nerve, respectively. These findings have led to the hypothesis that the pathogenesis of diabetic neuropathy is due to endoneurial hypoxia. To evaluate the role of endoneurial hypoxia in experimental diabetic neuropathy, the effects of supplementation and deprivation of oxygen on peripheral nerve lipid biosynthesis were studied in normal control and streptozotocin-induced diabetic rats. Defective lipid biosynthesis in diabetic nerve was partially prevented by oxygen supplementation. When normal rats were placed in a hypoxic chamber, lipid abnormalities similar to those observed in diabetic nerves were demonstrated in the absence of changes in nerve free sugars. These findings suggest that endoneurial hypoxia may underlie some key biochemical abnormalities encountered in experimental diabetic neuropathy.     

A clinico-pathological and follow up study of 10 cases of essential type II cryoglobulinaemic neuropathy.

Ten patients with essential cryoglobulinaemia type II were examined for peripheral nerve damage. In six cases distal symmetrical nerve involvement was present, while in three other cases abnormalities restricted to single nerves were found. Electrophysiological and morphological data were consistent with axonal damage, the larger myelinated fibres being most affected. Although active signs of vasculitis and immunoperoxidase staining for immunoglobulins were not present, endoneurial vessels were widely damaged, with abnormally thick endothelial cells and redundant basal membranes. These findings, together with a patchy distribution of myelinated fibre loss, suggest ischaemia as a cause of peripheral neuropathy during essential cryoglobulinaemia type II. A follow up examination, performed one year after haematologial remission, revealed that no further peripheral nerve damage had occurred.     

Peripheral nerve abnormalities related to galactose administration in rats.

Electrophysiological, biochemical, and morphometric observations were made on the peripheral nerves of rats after galactose feeding. Motor nerve conduction velocity was found to be reduced. This was associated with an accumulation of galactitol in the peripheral nerves and a diminution in their myoinositol content. An increased water content and fascicular area, taken in conjunction with a probable increase in the area of the endoneurial spaces, indicated overhydration of the peripheral nerves. Morphometric observations on the myelinated fibre population in the tibial nerve showed no loss of fibres and although both the maximal and the average diameter of the myelinated fibres was slightly less than in age-matched controls, this was insufficient to explain the reduction in conduction velocity. Segmental demyelination was not detected and the relationship between myelin thickness and axon circumference was not altered. Electron microscope observations revealed no ultrastructural changes in the myelinated fibres and, in particular, no abnormalities at the nodes of Ranvier or indication of abnormal hydration of the Schwann cells. The relevance of these findings to the peripheral nerve changes in human and experimental diabetes is discussed.     

Nerve lipid abnormalities in human diabetic neuropathy: a correlative study.

To study endoneurial lipid composition in human diabetic neuropathy, we biopsied sural nerves from 3 middleaged men with adult-onset diabetes mellitus. Magnitude of electrophysiological abnormalities and myelinated fiber loss paralleled the clinical severity of neuropathy in all cases. Cholesterol ester concentration was elevated to about 800% of normal in diabetic nerves. Reduction in total endoneurial lipid concentration correlated best with decrease in myelin volume as calculated from measured fiber diameters. Cholesterol, cerebroside, and most phospholipids were reduced in keeping with the severity of fiber loss in each nerve. The phosphatidylinositol-phosphatidylserine fraction was most reduced in the least affected nerves. Cerebroside nonhydroxy fatty acids in diabetic nerves were of shorter chain length and more saturated than normal. It is not yet clear whether the abnormalities of phosphatidylinositol-phosphatidylserine and cerebroside fatty acids are of pathogenetic importance or whether these changes may be the nonspecific consequence of axonal degeneration.     

Recent advances in the pathogenesis of diabetic neuropathy

Recent advances in the understanding of the pathogenesis of diabetic neuropathy have been made in six areas. There is support for the notion that a reduction in nerve free myoinositol may be responsible in part for the nerve conduction slowing in diabetic neuropathy. There is further evidence of microvascular abnormalities, including morphometric evidence of multifocal fiber loss and of capillary changes in biopsied sural nerve. There is evidence of endoneurial hypoxia, including the findings of reduced nerve blood flow and endoneurial oxygen tensions in chronic experimental diabetic neuropathy (EDN). The major mechanisms of resistance to ischemic conduction failure (RICF) is the marked increase in nerve energy substrates. Recent studies provide certain insights into clinical characteristics of human diabetic neuropathy (HDN), including the asymmetric pattern of HDN, the paradox between liability to pressure palsies and RICF, and insulin-related acute painful neuropathy. The suggested pathogenetic scheme incorporates the notion that once hypoxia is established, it may start a vicious cycle of further capillary damage and escalating hypoxia.     

A micro-electrode study of peripheral neuropathy in man. Part 1. Responses to single graded stimuli.

Multi-unit micro-electrode recording were obtained from sensory fascicles of the sural and median nerves of 12 control subjects and of 28 patients with peripheral neuropathy. Spontaneous activity and mass responses to mechanical and electrical stimuli were examined. Mechanoreceptor function appeared normal but there was a reduced number of responsive receptors in peripheral neuropathy. The electrical activation threshold of nerve fibres of all conduction velocities was increased in neuropathy and a greater number of fibres needed to be activated for preception to occur. Clinical sensory impairment was associated with a reduction in size of the initial compound action potential of the maximal evoked neurogram and with dispersion of fibre responses. Pathological slowing of fibre conduction velocity was demonstrated in demyelinating neuropathy but in most cases of axomal degeneration the changes in velocity could have been due either to a reduced number of fast conducting fibres, or to conduction block. No changes were observed in C-fibre activity in these patients.     

Peripheral nerve pathological findings in familial amyloid polyneuropathy: a correlative study of proximal sciatic nerve and sural nerve lesions

To analyze the peripheral nerve pathological abnormalities in familial amyloid polyneuropathy, a correlative pathological study was carried out on the spinal nerve roots, proximal sciatic nerves, sural nerves, and brachial plexuses from 3 patients with the disease in Japan. The spinal nerve roots appeared to be unaffected except for amyloid deposition on the epineurium. In sciatic nerves and brachial plexuses the nerve lesions had a multifocal distribution, showing prominent interstitial edema in the endoneurium frequently adjacent to deposits of amyloid; in these regions the nerve fibers were severely depleted. A teased-fiber study revealed that segmental demyelination was the predominant type of nerve fiber abnormality. However, these findings were not seen in the sural nerves; instead a diffuse fiber loss with axonal degeneration was observed. It is suggested that multifocal lesions in the proximal portions of the long extremity nerves could summate distally to produce a symmetrical polyneuropathy in the disease. In addition to a space-occupying effect of amyloid deposits in the endoneurium, severe endoneurial edema associated with amyloid deposition in blood vessels and the endoneurial interstitium may induce ischemia in nerve fibers, thus causing the progressive polyneuropathy in this disorder.     

Endoneurial microvascular abnormalities of sural nerve in non-diabetic chronic atherosclerotic occlusive disease

Neuropathic abnormalities are found in chronically ischaemic limbs associated with non-diabetic atherosclerotic peripheral vascular disease (PVD). In chronic ischaemic neuropathy, microvascular alterations play a key role in its development. We undertook morphometric assessment of endoneurial microvessels in the sural nerves, taken from severely ischaemic amputated legs in nine chronic non-diabetic PVD. These subjects had threatened ischaemic limbs and revealed clinical, physiological and pathological evidence of neuropathy. For comparison, sural nerves taken from amputated legs due to non-ischaemic disorders (n=4) and chronic PVD associated with diabetes (n=3) were also assessed. We evaluated the areas of vascular lumen, endothelial cells and whole vessel, as well as the percentage of closed capillaries. Endothelial area of sural nerve microvessels in non-diabetic PVD nerves was significantly greater than in non-ischaemic control nerves. Periendothelial cell area containing pericytes and basement membranes was also significantly increased in non-diabetic PVD nerves when compared with control nerves. Vascular lumen area was significantly less in non-diabetic PVD nerves than in non-ischaemic control nerves. Endoneurial microvessels in diabetic PVD nerves showed similar results: thickened vessel wall and smaller lumen. Periendothelial area in diabetic nerves was significantly greater than in non-diabetic PVD nerves. We demonstrated swollen endothelial cells and increased periendothelial area associated with narrowed lumen in sural nerve endoneurial microvessels of severe chronic PVD. Basement membrane reduplication of endoneurial capillaries was seen in non-diabetic PVD nerves. These microvascular abnormalities could play an important role in the development of chronic ischaemic neuropathy in PVD limbs.     

Friedreich's ataxia: electrophysiological and histological findings

ABSTRACT- Electromyography was performed, and motor and sensory nerve conduction velocities were measured in 19 patients definitely affected by Friedreich's ataxia. Biopsy of the sural nerve was also performed in 9 patients.
Most patients presented a moderate to severe loss of motor units, a significant increase in mean duration of motor unit potentials, and in the incidence of polyphasic potentials. Short-lasting spontaneous activity was rarely seen. Conduction velocity along the motor and sensory fibres of the median and tibial nerves was moderately slowed, while distal conduction time to muscle was significantly increased and the sensory orthodromicallyevoked response markedly reduced. Intraoperative electrophysiological recordings obtained during biopsy of the sural nerve in 4 patients were consistent with the changes conventionally observed in the median, tibial and sural (6 patients) nerves.
Quantitative histology revealed a reduced number of total myelinated fibres with a severe loss of large fibres, and a moderate loss of fibres of less than 7 μm in diameter. In teased nerve fibre preparations, the most evident abnormality consisted of fibres with uniformly short internodal length, while signs of remyelination were less prominent. Signs of active axonal degeneration were rarely observed in electron microscopy.
Electrophysiological and histological findings were uniformly distributed, and the changes were neither related to the duration nor to the severity of the clinical condition.     

Correlation between the number of epineurial and endoneurial blood vessels in diseased human sural nerves

In a preceding study it was shown that changes in the number of epineurial blood vessels may be a prominent feature in angiopathic and other peripheral neuropathies, for instance in vasculitis, diabetes mellitus, or cerebral autosomal dominant angiopathy with multiple infarcts and leukoencephalopathy (CADASIL). Endoneurial blood vessels usually may also show significant structural alterations in a broad spectrum of neuropathic conditions, although these are not as prominent as in the epineurium. However, the relationship between changes in the number of epineurial and endoneurial blood vessels in diseased human sural nerves, and the impact of the loss of myelinated nerve fibers on the number of endoneurial blood vessels has thus far not been determined. Therefore, we investigated and compared the number of epineurial and endoneurial blood vessels in 50 human sural nerve biopsy specimens, representing a variety of peripheral neuropathies. We found that despite a significant increase of the number of epineurial blood vessels in cases with vasculitic neuropathy (P<0.05) and neuropathy with other types of microangiopathy (P<0.01), the number and density of the endoneurial blood vessels remained remarkably constant. In cases with an axonal type of neuropathy, severe neuropathic changes were associated with a decreased epineurial blood vessel number and a simultaneous, relative increase in the endoneurial blood vessel density. No significant correlation was found between (1) the number of epineurial and endoneurial blood vessels, and (2) the severity of the neuropathy and the number or density of epineurial and endoneurial blood vessels.     

C-Peptide Deficiency: An Important Pathogenetic Factor In Type 1 Diabetic Neuropathy

12 diabetic patients aged 47.5 ± 9.4 yr., duration of diabetes (14.6 ± 10.3 yr.) and 15 control subjects were studied. In diabetic patients neuropathy symptom score =0, neuropathy deficit score = 4.5 + 0.7/30, vibration = 12.0 + 1.8 V, thermal perception (2.0 + 0.8°C), heart rate variation during deep breathing (17.8 + 2.3), 30:15 ratio (1.31 + 0.07) was normal. Baseline (n=12) and repeat neurophysiology (n=10) performed 8.7 + 0.6 years after sural nerve biopsy demonstrated normal values at baseline, with progression of neuropathy (peroneal motor nerve conduction velocity (ms⁻¹) (42.3 + 2.9 v 39.4 +2.0), sural nerve conduction velocity (45.4 + 3.7 v 43.6 + 1.7). Myelinated fibre density, fibre and axonal area and g-ratio were not significantly reduced. Teased fibre studies showed paranodal abnormalities (p < 0.001), segmental demyelination (P < 0.01) with remyelination (P < 0.01) without axonal degeneration. Unassociated Schwann cell profile density (p < 0.04) and axon density (P < 0.001) were increased and axon diameter was decreased (P < 0.007) with a shift of the size frequency distribution to the left (skewness- 0.89 v 0.64, P < 0.03) suggestive of unmyelinated axonal atrophy/regeneration. Endoneurial capillary basement membrane thickening (P < 0.006), endothelial cell hyperplasia (P < 0.004) and luminal narrowing (P < 0.007) occurred. Current measures of neuropathy are too insensitive to detect significant nerve fibre pathology. The presence of microangiopathy provides support for a microvascular basis of diabetic neuropathy.     

Unmyelinated fibres and Schwann cells of sural nerve in neuropathy.

Electron micrographs of 45 sural nerves from patients with acquired (22) or heredodegenerative neuropathy (23) were analysed with respect to the number of unmyelinated nerve fibres, 37 nerves with respect to the number of Schwann cell sub-units and of structures connected with Schwann cells. Findings were compared with those in 6 nerves from control subjects and referred to the total number rather than to the number per mm2 to eliminate error due to increase in the transverse endoneurial area, present in more than half the diseased nerves. Ninety-one per cent of the diseased nerves showed one or several abnormalities in unmyelinated fibres of their Schwann cells. The best indicator of fibre loss was an increase in the number of Schwann cell sub-units devoid of axons, found in more than half the nerves. This was the only abnormality related with decrease in number of myelinated fibres. The increase in number of empty Schwann cell sub-units was due both to loss of unmyelinated nerve fibres and to proliferation of Schwann cells. Proliferation was indicated by the higher incidence of Schwann cell nuclei in cross-sections of diseased nerves than in controls. The earliest sign of involvement was an increase in number of profiles and of small isolated Schwann cell projections, observed in 33 of 37 diseased nerves, as the only abnormality in 7 nerves. The number of unmyelinated nerve fibres by itself was of little value to indicate loss of fibres, since regeneration often replaced or more than replaced degenerated fibres. Regeneration was indicated by an increase in number or incidence of small unmyelinated fibres, present in nearly half of 45 diseased nerves; and by an increased in the total number, present in a third of the nerves. An increase in the number of collagen pockets and of fibres undergoing degeneration (loss of organelles) and a decrease in the number of unmyelinated fibres per Schwann cell sub-units was present in only a quarter to a third of diseased nerves and was not related to other criteria of loss of fibres or of regneration.     

Prevalence of subclinical neuropathy in diabetic patients: assessment by study of conduction velocity distribution within motor and sensory nerve fibres

Nerve conduction velocity distribution (CVD) study is a newly-developed electrodiagnostic method for detecting alterations in the composition of nerve fibres according to their conduction velocity. The presence of subclinical neuropathy was evaluated in 138 diabetic patients by CVD study of four motor nerves (external popliteal and ulnar nerves bilaterally) and two sensory nerves (median nerve bilaterally), and the data obtained were compared with standard electrophysiological parameters in the same nerve segments. CVD studies revealed an altered distribution pattern in 106 of 129 evaluable patients for motor nerves (82%) and in 67 of 115 evaluable patients for sensory nerves (58%), while standard examination gave abnormal findings in 92 of 137 patients (67%) and in 33 of 118 patients (11%), respectively. Of the patients adequately evaluated by both techniques, 21 of 129 patients (16%) revealed altered CVD data unaccompanied by slowing of maximum nerve conduction velocity, and 37 patients of 101 (37%) showed similar findings for sensory nerves. Subclinical alterations of motor and sensory nerve CVD were not significantly related to age or to metabolic control expressed as glycated haemoglobin levels; a significantly longer duration of disease was found in patients with motor and mixed subclinical neuropathy with respect to non-neuropathic patients. The CVD study allowed us to detect subclinical abnormalities of motor and sensory nerve fibres; often this is a more sensitive method than the standard electrodiagnostic study. This method can be very useful as a diagnostic tool and in research in the study of the progression of diabetic neuropathy. Received: 21 March 1997 Received in revised form: 8 September 1997 Accepted: 7 October 1997     

Alcoholic neuropathy: Clinical,electrophysiological, and biopsy findings

Nerve conduction and biopsy findings from the sural nerve in 37 patients with alcohokic neuropathy were compared with findings in 6 patients who had neuropathy associated with postgastrectomy malnutrition. Half the patients with alcoholic neuropathy had both muscle weakness and sensory loss, half had only sensory impairment, but all had electromyographic signs of denervation. Only half the patients, with or without muscle weakness, had signs of malnutrition. In alcoholics, sural nerve conduction velocity was slowed to at most 60% of normal, correlating with loss of large fibers. These findings, together with a marked reduction in amplitude of the sensory potentials, are consistent with axonal loss. Myelinated fiber counts showed loss of small and large fibers in most nerves, retaining a bimodal distribution. Signs of regeneration were rare. Segmental demyelination was found in only 0.3% of teased fibers. Electron microscopy confirmed axonal degeneration of myelinated and unmyelinated fibers. Neuropathy after gastrectomy malnutrition was clinically similar to alcoholic neuropathy. Conduction velocities were slower than expected from the diameter of the largest myelinated fibers, however, and teased fibers showed segmental demyelination. The findings are against alcoholic neuropathy being due to malnutrition and suggest a toxic action on peripheral nerve.     

Leprous neuropathy: a clinical and neurophysiological study

Leprosy is one of the most common treatable causes of neuropathy in the world. Peripheral nerves and skin are commonly affected. We reported the clinical features and electrophysiological findings in 46 patients with leprosy. The aim of our study was to evaluate the nature of damage in the nerve fibres, especially in the first phase of disease. Forty‐six patients (mean age: 44.8 ± 17.8) with diagnosed leprosy were studied by neurological examination and nerve conduction studies (NCS). Twenty‐eight patients were examined for a mean period of 34.8 months. The number of tests for patients varied from 1 to 13 controls. Amplitude of sensory and motor action potentials (SNAP and MAP), sensory‐motor conduction velocity of median, ulnar, tibialis, peroneal and sural nerves were evaluated. Abnormalities were found in 282 of 647 nerves investigated (37.56%), sensory nerve abnormalities being more frequent than motor (50.16% 29.45). Of 282 nerves with neurophysiological abnormalities, 123 were clinically asymptomatic (43.62%). A statistically significant correlation between duration of disease and number of electrophysiological abnormalities was demonstrated. In 19 nerves partial "conduction‐block"(reduction of cMAP > 50% in the proximal response) was individuated. The first electrophysiological alteration, suggesting segmental demyelination, was detected in 41 nerves of 21 patients (33.3 %). According to this view, our data support the hypothesis that leprosy induces a neuropathy of demyelinating nature in the first phase.     

Myxoedematous polyneuropathy: a light and electron microscopic study of the peripheral nerve and muscle.

Histopathological findings of biopsied peripheral nerve and muscle were studied in a case with myxoedematous polyneuropathy. The most striking findings in the sural nerve were segmental demyelination and onion bulb formation with scanty mucinous deposits in addition to marked loss of large myelinated nerve fibres. The peroneus brevis muscle revealed the association of neuropathic and myopathic changes. It is suggested that myxoedematous polyneuropathy might be intrinsic neuropathy due to metabolic disorder of Schwann cells related to hypothyroidism, resulting in segmental demyelination, not merely compressive neuropathy due to mucinous deposits in the peripheral nerves.