While individual cases of CMT-1A and CMT-1B are not clinically distinguishable, CMT-1B is typically more severe (Dyck et al. 1993). Symptoms usually begin in the first or second decade although nerve conduction is abnormal in early life. Onset in the 3rd–5th and rarely up to seventh decade is well described (Harding and Thomas 1980a). The usual presenting manifestations are skeletal deformity, distal leg muscle atrophy, weakness, or indirectly through findings detected while undergoing evaluation for non-neurological symptoms (Dyck et al. 1993). Pes cavus, hammertoes, and scoliosis may be present as early manifestations in childhood and serve as important clues to the diagnosis of a familial neuropathy. Symptoms can be so mild and insidious that some patients never present to a physician for disease-related complaints. Visible and palpable hypertrophy of peripheral nerves, especially the greater auricular nerve, can be found in approximately a quarter of patients. Weakness and atrophy are often restricted to the distal legs, even after many years of illness, with about half of patients progressing to distal hand muscle weakness. Reflexes are typically diminished or absent in a distal to proximal gradient. Sensory loss involves all modalities but is usually less prominent than motor symptoms, joint destruction, or sensory ataxia being uncommon. Essential tremor is seen in a third of CMT-1 patients and its presence defines the “Roussy–Lévy” variant, but these patients are not genetically distinct from those with other CMT-1 variants. The disease progresses very slowly over a patient’s lifetime and does not influence most individuals’ longevity; only a minority eventually become wheelchair dependent. A small subset of patients with dominantly inherited CMT-1 present in early childhood or even at birth, some progressing very slowly, but others reaching a debilitated state (Ouvrier et al. 1987; Vanasse and Dubowitz 1981; Hagberg and Lyon 1981).

Medication-induced (vincristine, metronidazole, nitrous oxide, nitrofurantoin, etc.) exacerbation of neuropathy in CMT is well known (Weimer and Podwall 2006). In CMT patients, acute initiation of weakness, including a Guillain–Barré-like pattern, may occur. In ten previously unsuspected and undiagnosed CMT patients, nerve deficit developed following the administration of pharmaceuticals as the initial manifestation of the genetic disease.

Seven different defects in the gene encoding the gap junction protein connexin-32 have been detected in 8 families with X-linked CMT (Bergoffen et al. 1993b), while unaffected family members and control subjects did not possess the mutations. This type of CMT disease, which is probably the second most common form of the disease (accounting for around 10 % of all CMT patients), is a genetically separate, X-linked condition. More than 250 mutations of the CX32 (connexin 32, gap junction protein β1, GJB1) gene have been found to be the cause of this syndrome (Bergoffen et al. 1993a, b). Males with CMT-X display a more severe clinical, electrophysiological and histopathological disease than heterozygous females, befitting the notion that having two X chromosomes confers some protection. Connexin-32 is a member of a family of transmembrane proteins, the connexins, which assemble to form gap junctions. Immunohistochemistry reveals that these proteins are found in myelinated peripheral nerve fibers at the nodes of Ranvier and at Schmidt–Lanterman incisura (Bergoffen et al. 1993b). The workers identifying this mutation hypothesized that dysfunction of putative intracellular gap junctions at these sites might impair flow of ions and nutrients to the innermost layers of compact myelin and perhaps the axon, resulting in myelin disruption and axonal degeneration (Bergoffen et al. 1993a, b).

The pathological changes are somewhat nonspecific, but just as the condition is often considered “intermediate” between CMT-1 and CMT-2, so too are its nerve conduction velocities and histopathological findings (Sander et al. 1998; Vital et al. 2001). Classic onion bulbs are not very frequent, but “pseudo-onion bulbs” are common. In these formations, a regenerating cluster is centered by a myelinated fiber and surrounded by several unmyelinated axons and their associated Schwann cells. Axon numbers are often relatively preserved in CMT-X, as large myelinated fibers degenerate or atrophy, but regenerative clustering activity provides numerical compensation. As is typical for genetic processes, the histopathological picture appears chronic, with no active axonal loss (i.e., Wallerian degeneration) or active demyelination visible on cross sections, although teased fibers may show some ongoing axonal degeneration.

When a nerve biopsy shows prominent hypertrophic changes, the most important diagnostic considerations are CIDP and CMT-1. These can usually be distinguished by the pattern of involvement: multifocal in CIDP and diffuse in CMT-1 and definitively by genetic analysis. However, minimal pathological expression of CMT-1 may demonstrate multifocal regions containing onion bulbs interspersed with fibers of normal appearance (Dyck et al. 1983). Prominent inflammatory infiltrates are not typical of classic CMT-1; however, a few epineurial lymphocytes should not be over interpreted as CIDP. Because of the chronicity of the process, macrophages filled with myelin debris are rare in adult CMT-1, and when several are seen in the same section, CIDP should be considered. We do not accept macrophage-mediated myelin stripping as a feature of CMT-1. If evidence is conflicting, consideration should be given to the possibility of CMT with a superimposed inflammatory neuropathy, which may be steroid responsive (Dyck et al. 1982; Bird and Sladky 1991; Vital et al. 1990; Crawford and Griffin 1991). Gabreels-Festen et al. (1993) have considered this issue, and a detailed discussion is provided in Chap.​ 9. PMP-22 expression in nerve biopsy material can distinguish between CMT-1A and other demyelinating neuropathies, but there is an overlap between the two groups (Yoshikawa et al. 1994).

In an infant with a dysmyelinating neuropathy, the differential diagnostic possibilities include CMT-1, Dejerine–Sottas syndrome, the other congenital dysmyelinating neuropathies (vide infra), and CIDP. An infantile or early childhood neuropathy progressing for many years can be due to CIDP, and nerve biopsy may be the only means of making the diagnosis (Sladky et al. 1986), although nerve conduction studies are very helpful. In the absence of a family history of dominant inheritance, sporadic infantile CMT-1 is distinguished from Dejerine–Sottas in that the latter shows very prominent hypomyelination and demyelination with a g-ratio >0.7, few or no axons over 6–8 μm in diameter on the myelinated fiber histogram, and prominent (>50 % of myelinated fibers) onion bulb formations at a young age (Ouvrier et al. 1987). The distinction is important because prognosis is better for infantile onset CMT-1 than for Dejerine–Sottas syndrome (Vanasse and Dubowitz 1981; Ouvrier et al. 1987). Moreover, nerve biopsy might suggest the mode of transmission, i.e., dominant (CMT-1) or recessive (Dejerine–Sottas). The available literature suggests that biopsies revealing numerous basement membrane onion bulbs or focally folded myelin are more likely to be recessive (vide infra). Uniformly thin myelination is not specific for Dejerine–Sottas syndrome. We have examined CIDP biopsies in adults with a picture of uniformly distributed, strikingly thinly myelinated axons in an adult, and a similar appearance may be present in metachromatic leukodystrophy (Bardosi et al. 1987). Nerve biopsy features typical of Dejerine–Sottas syndrome were seen in a patient with a point mutation in the PMP gene who had dominant inheritance of a particularly severe CMT-1 phenotype (Hoogendijk et al. 1993) and in a patient with homozygous inheritance of dominant CMT-2 (Sghirlanzoni et al. 1992).

CMT-1 can show dramatic histological variability. In one CMT-1 kinship in which 5 biopsies were performed, one showed classic OBs, 2 showed thinly myelinated axons, and 2 showed only axonal loss (van Weerden et al. 1982).

Hereditary neuropathy with pressure palsies (HNPP) is an increasingly recognized entity that has emerged in recent decades. Patients experience recurrent bouts of mononeuropathy, sometimes precipitated by trivial injury to the affected nerve and usually resolving over hours to weeks. Painless peroneal, radial, or ulnar palsies are most common (Meier and Moll 1982; Verhagen et al. 1993; Earl et al. 1964), and a mononeuritis multiplex-like picture may occur (Behse et al. 1972). Brachial plexus involvement is not infrequent, accounting for 8 % of episodes in one review (Meier and Moll 1982), while cranial nerves are only rarely affected. A mild polyneuropathy may be superimposed. Uncommonly, weakness in the territory of a single nerve can be slowly progressive but may remit even after many years if close attention is paid to avoidance of pressure or traction (Earl et al. 1964, case 19-1). Although most patients identify the second or third decades as the time of disease onset, focal paralysis may be present at birth. Alternatively, the diagnosis can be delayed as long as the seventh decade if the recurrent nature and family history of such symptoms are not appreciated (Meier and Moll 1982). The inheritance pattern is autosomal dominant, with complete or nearly complete penetrance but variable disease severity (Verhagen et al. 1993). Genetic studies have shown point mutation in PMP22 or deletions affecting the region of chromosome 17 harboring the PMP22 gene (Chance et al. 1992, 1993) in 85 % of patients. Therefore, HNPP can be caused by deletion of the gene PMP22 while duplication of PMP22 causes Charcot–Marie–Tooth disease type 1A, and point mutation in PMP22 may result in either HNPP or CMT1A.

Rarely, patients with the typical histological findings of HNPP seem to suffer a slowly progressive, distally predominant sensorimotor polyneuropathy (Madrid and Bradley 1975; Malandrini et al. 1992; Barbieri et al. 1990; Pellissier et al. 1987) similar to the Charcot–Marie–Tooth syndrome. It is unclear if this is a phenotypic variant, a spurious consequence of insufficient detailed electrophysiological testing (Felice et al. 1994), or an entirely different disease, perhaps CMT-1B (Thomas et al. 1994). The pediatric cases of HNPP reported by Gabreels-Festen and colleagues (1992a) often had a chronic progressive neuropathy, while a history of recurrent focal palsies could be found in some of the affected older relatives, suggesting that HNPP can manifest as a chronic progressive neuropathy in early life, with the typical focal palsies coming later. Other pedigrees exist in which some affected individuals showed recurrent focal palsies while others had only a chronic neuropathy (Leblhuber et al. 1991).

Electrophysiological testing reveals a diffuse abnormality of peripheral nerves, with mild-moderate slowing of conduction, worse distally, even in clinically unaffected nerves. Sites of nerve compression are particularly likely to be affected. EMG reveals neurogenic changes, usually chronic, indicating that axonal degeneration does occur (Behse et al. 1972). CSF examination is unremarkable.

Madrid and Bradley (1975) have discussed the mechanisms that might form the myelin swellings of HNPP, including infoldings, outfoldings, and splitting of the mesaxon. These authors pointed out that similar findings are seen during myelinogenesis but that ordinarily myelin remodeling produces the normal smoothly contoured sheath. The morphologic findings of HNPP are not disease specific, but can be seen, albeit less dramatically, in other demyelinating diseases. A reasonable postulate is that there is a defect in either the ability of the axon to send regulating signals to the Schwann cell or a defect in the Schwann cell’s ability to interpret signals from the axon.

Yoshikawa and Dyck (1991) observed uncompacted myelin in HNPP and found evidence that this change could precede segmental demyelination. The suggestion was made that the neuropathy might arise due to a defect in membrane proteins or lipids involved in myelin compaction, P₀ protein and myelin-associated glycoprotein being identified as leading candidates. Gross chemical and immunohistochemical analysis of the myelin produced in HNPP reveals no abnormalities of composition, and Schwann cell cultures from patients with HNPP showed no special features when compared to controls (Federico et al. 1989).

In 1993 a chromosome 17 deletion was found to be strongly associated with HNPP in 3 pedigrees (Chance et al. 1993), and this observation has subsequently been confirmed in both inherited and sporadic cases of HNPP (Verhalle et al. 1994; Roa et al. 1993b; Reisecker et al. 1994). Remarkably, the deleted region is seemingly identical to that normally duplicated in CMT-1A (vide supra), such that only one copy of the PMP-22 gene is present in patients with HNPP. Although attention has focused on PMP22, other genes in this deleted region or genes disrupted by the deletion could contribute to the problem. Mutation in the myelin P₀ protein has recently been reported to cause a neuropathy with clinical features of CMT-1 and histological features of HNPP (Thomas et al. 1994).

Focal myelin swelling, excessive myelin infolding and outfolding, and circumferential hypermyelination are in and of themselves nonspecific, occurring to varying degrees in classic CMT-1, CMT-3, congenital dysmyelinating neuropathy, IgM paraprotein-associated neuropathy, and neuropathy with unstable myelin. Such alterations can also be present to a limited degree in patients with nonspecific neuropathy or even without neuropathy (Drac 1989). However, in the large HNPP series reviewed, above no less than 5 %, and usually a greater proportion, of internodes contained tomaculae (Drac 1989). This is beyond the frequency of such changes in patients with classic CMT-1 or Dejerine–Sottas syndrome (vide infra) or in controls (Drac 1989). No diagnostic significance can be attached to a nerve biopsy with myelin swellings or foldings on less than 1 % of internodes (Drac 1989).

A growing number of patients and pedigrees have been described where histological findings typical of tomaculous neuropathy were found in association with an atypical history: suggestive of either CMT (neuronal or hypertrophic) (Malandrini et al. 1992; Madrid and Bradley 1975; Barbieri et al. 1990; Gabreels-Festen et al. 1992b; Thomas et al. 1994) or even GBS (Joy and Oh 1989). It is unclear if these represent phenotypic variants of the same disease or if they are altogether different entities with tomaculum formation as a nonspecific response. The severity of axonal loss described in some of these reports is more typical of CMT-1 than of HNPP, and a mutation in the myelin P₀ protein has been identified in one instance (Thomas et al. 1994), suggesting that at least some of these cases represent HSMN-1B rather than HNPP.

The patients designated by Gabreels-Festen and Gabreels (1993) as “CMT-1 with focally folded myelin” have myelin tomaculae that resemble those of HNPP in size, ultrastructural appearance, and frequency (vide supra). However, hypomyelination, onion bulb formation, and especially axon loss are much more prominent in this group than in HNPP. Clinically there should be little trouble distinguishing the two, as in “CMT-1 with focally folded myelin” focal palsies are not a feature, the phenotype and electrophysiological findings are more severe, and autosomal dominant inheritance is not usually seen.

In IgM paraproteinemic neuropathy, typically associated with antibody activity against myelin-associated glycoprotein, redundant myelin folds and hypermyelination very similar to that seen in HNPP are sometimes seen (Vital et al. 1989). Widely spaced, not uncompacted, myelin is present in IgM paraproteinemic neuropathy, allowing a distinction to be made purely on morphologic grounds, although clinically there is no difficulty separating the two.

Frequent “sausage-like” myelin swellings apparently typical of HNPP have been described by Said and colleagues in alcohol-associated “acrodystrophic” neuropathy (Said 1980), uremic neuropathy (Said et al. 1983), and in cytomegalovirus neuritis in a patient with AIDS (Said et al. 1991). However, the ultrastructural correlate of these swellings was not provided, other workers have not reported such data, and we have never made this observation in a number of nerves taken from patients with these diseases.

We wish to emphasize that “tomaculous neuropathy” is not synonymous with HNPP. If typical histological features of the tomaculae are present, HNPP can be diagnosed with a high degree of confidence, although a small group of patients with a chronic progressive neuropathy will remain whose classification is uncertain at present. Molecular genetics should help resolve this issue. However, detection of rare myelin tomaculae or redundant myelin loops is of little diagnostic utility.

Autosomal dominant inherited tendency to recurrent brachial plexopathy (inherited brachial plexus neuropathy = IBPN) has been described (Windebank 1993). Symptoms of the individual attack are identical to those of idiopathic brachial neuritis. Pain, which can be very severe, heralds the onset of paralysis by hours to days. Weakness occurs predominantly in a proximal upper limb distribution but can be present in the distal arm, cranial nerves, and lower limbs. Recovery is the rule, although this may take many months. Nosologically, this entity poses a problem, for it shares many features with HNPP. Patients with HNPP may have episodic weakness involving brachial plexus muscles, and patients with IBPN can involve the lower limbs (Dunn et al. 1978). Compression or traction on the brachial plexus may precipitate an attack in both entities. Inheritance is autosomal dominant and typical age of onset is in the second or third decade for both. Nevertheless, most neurologists separate the two based on clinical and electrophysiological criteria (Windebank 1993; Verhagen et al. 1993), although there are dissenters (Martinelli et al. 1989). Pain with paralysis is an almost invariable feature of IBPN, and preceding infection and pregnancy are common precipitating factors. Neither of these is true for HNPP. In HNPP electrophysiological tests reveal evidence of a diffuse sensorimotor polyneuropathy (Behse et al. 1972). In IBPN most studies have pointed to isolated involvement of the plexus (Windebank 1993), but not all have found this (Dunn et al. 1978).

There is little histological material on IBPN, presumably because it does not seem rational to biopsy the sural nerve in a disease of the brachial plexus. In one case studied at the Mayo clinic, no abnormality was found on exploration and fascicular biopsy of the brachial plexus (Windebank 1993). Verhagen et al. (1993) make reference to two patients in which sural nerve biopsy similarly did not reveal myelin tomaculae. Arts et al. (1983) found no myelin tomaculae in nerve biopsy and autopsy of 2 patients with IBPN. This data would suggest that IBPN is a different entity with different histological findings than HNPP.

Several pedigrees have been described with familial recurrent brachial plexopathy and myelin tomaculae on nerve biopsy (Martinelli et al. 1989; Pou Serradell et al. 1992). These patients had a painless disease, showed diffuse abnormalities on electrophysiological testing, and are thus likely to represent examples of HNPP with selective involvement of the brachial plexus. However, case 1 reported by Madrid and Bradley (1975) had prominent pain with her illness and tomaculae on nerve biopsy, as did the case we report below. The IBPN locus had been mapped to chromosome 17q24-25, and recently, genetic analysis has shown that IBPN is caused by mutation in the SEPT9 gene (Hannibal et al. 2009) which encodes a septin family member, a family of genes most involved in cytokinesis and cell cycle control. However, there remains a proportion of families with IBPN without the genetic abnormalities described above, indicating the existence of other genes involved.

Neuronal (type 2) CMT (CMT-2) is the typically autosomal dominant (50–70 %) neuronal form of the disease showing motor nerve conduction velocities of 38 m/s or greater, with 25–50 % sporadic and 5 % recessive cases in the largest series (Harding and Thomas 1980c; Bouche et al. 1983; Berciano and Combarros 1990). Segregation analysis suggests that 25 % of sporadic cases are recessive and the rest new dominant mutations (Harding and Thomas 1980c). Genetic analysis of autosomal dominant CMT-2 now has identified 19 gene defects and 5 abnormal genes in autosomal recessive CMT-2. Its prevalence is about one-third that of CMT-1 and CMT-2A; its most frequent form (20 % of CMT-2 cases) is associated with a missense mutation of Mfn2 (mitochondrial fusion protein 2 gene). Mitofusin (Mfn2) is a mitochondrial membrane protein that participates in mitochondrial fusion in mammalian cells and in pyruvate, glucose, and fatty acid oxidation, defects of which culminate in reduced mitochondrial membrane potential. Experimental data support the notion that Mfn2 loss of function causes axonopathy by impairing energy production along the axon which may reflect abnormal tethering of mitochondria to ER (Pich et al. 2005). Type 2A CMT is clinically similar to CMT-1, although on average the phenotype is milder (peripheral nerves are not enlarged) and the disease develops later (mean age of onset 20 years), there is less prominent involvement of upper limbs (particularly the hands) and less ataxia, tendon areflexia, and sensory loss. Mutations in the RAB7 (small GTPase late endosomal protein) gene cause CMT-2B, a predominantly sensory neuropathy complicated by chronic foot ulcers and amputations (Lawson et al. 2005). Type CMT-2D is associated with mutations of the GARS (glycil tRNA synthetase) gene and consists of upper extremity wasting and weakness, with corresponding sensory impairment; lower extremities are involved to a lesser degree (Lawson et al. 2005). Mutations in the NEFL gene are associated with axonal neuropathy and designated as CMT-2E. This condition is characterized by variable severity and sensory ataxia. NCS may show preservation of normal values or slight slowing. Pathological findings are of interest in that they reveal the presence of giant axons with attenuated myelin and containing aggregates of disorganized neurofilaments (Pareyson et al. 2013).

Autosomal recessive forms of CMT-2 are called autosomal recessive CMT (AR-CMT). Recessive mutations of the LMNA (laminin A/C) gene are the cause of the axonal neuropathy in AR-CMT-2 (formerly known as CMT-4C and CMT-2B1). Whether a neuronal X-linked form exists is unclear (Hahn 1993; Dyck et al. 1993). Clinical manifestations are much like those of CMT-1, summarized above. No clinical features definitively distinguish the two, although onset in the first decade and palpable nerves suggest the hypertrophic form, and variations in distribution of atrophy, weakness, and reflex changes may be helpful (Hahn 1993). As with CMT-1, examination of “normal” family members is essential because the disease may be asymptomatic.

Patients with childhood onset are more severely affected, often resulting in debilitation by adulthood (Ouvrier et al. 1981; Gabreels-Festen et al. 1991). Most of these are probably genetically distinct from dominantly inherited CMT-2. From a total of 29 cases with childhood onset and a severe clinical course (Gabreels-Festen et al. 1991; Ouvrier et al. 1981), only 3 had a dominant inheritance pattern, 8 had similarly affected siblings and normal parents suggesting recessive inheritance, and most were sporadic.

Nerve conduction studies are usually the basis for a definitive distinction of CMT-2 from CMT-1, but rarely there are situations where the conduction velocities are slowed, presumably due to the selective and sometimes very severe loss of large (rapidly conducting) myelinated fibers (vide infra). Conduction velocity is usually normal or near normal, with reduced amplitudes suggesting loss of axons. EMG shows changes of very chronic denervation. CMT-2 has been reviewed by Hahn (1993), Dyck et al. (1993), and genetic analysis (Rossor et al. 2013; Saporta and Shy 2013; Sagnelli et al. 2013; Tazir et al. 2013; Vallat et al. 2013).

There are several large CMT-2 biopsy series (Behse and Buchthal 1977; Gherardi et al. 1983; Ouvrier et al. 1981; Berciano et al. 1986; Gabreels-Festen et al. 1991).

The genetics of the various forms of CMT-2 (Rossor et al. 2013) suggest a variety of pathogenetic mechanisms ranging from altered mitochondria and mitochondria-based metabolism of axons/Schwann cells to axonal cytoskeleton, endosomal functions, and others. Autopsy and teased fiber studies support the hypothesis that a distal axonopathy with axonal atrophy is the fundamental pathogenic mechanism (Berciano et al. 1986; Yasuda et al. 1990; Dyck et al. 1993).

Individually, the histological findings of CMT-2 are nonspecific. However, a biopsy showing diffuse loss of myelinated fibers, with relatively selective large fiber involvement, no active degeneration, and very prominent regenerating cluster formation, is highly suggestive of the diagnosis. In theory, a similar picture may be seen in other very longstanding axonal neuropathies such as diabetes, chronic renal failure, or long duration toxic exposures, but in our experience these always show a degree of active (Wallerian) degeneration that is out of keeping with the diagnosis of CMT-2.

Nerve biopsy shows large numbers of onion bulbs, endoneurial fibrosis, loss of axons, and, specifically, hypomyelination: Many fibers that, on the basis of axon diameter, should be myelinated lack myelin completely or show disproportionately thin myelin sheaths (Fig. 19.14a, b) (Guzzetta et al. 1982). Histological features are reminiscent of dominant CMT-1, but more severe as regards nerve enlargement, onion bulb formation, and axonal loss (Fig. 19.14). As many as 50 % of internodes may be demyelinated on cross sections and even more dramatically on teased fibers (Dyck and Gomez 1968; Dyck et al. 1970, 1971b; Ouvrier et al. 1987). Despite the large number of denuded axons, macrophages containing myelin debris are not prominent. The most salient observation, however, is that when axons do have a myelin sheath, it is always inordinately thin, regardless of axon size (Figs. 19.14 and 19.15). The myelinated axon population is reduced, sometimes very severely, with a marked shift towards small diameters. There can be overlap between the degree of axon loss seen in autosomal dominant CMT-1 and DSS in patients of roughly the same age (Ouvrier et al. 1987). Unmyelinated axon counts seem normal, but with the shift of the myelinated axon population towards small fiber size and the florid demyelination, it may be difficult to classify small nonmyelinated axons.

Onion bulbs are a prominent feature and may reach large proportions. The lamellae are made up of concentric imbricated layers of Schwann cell processes, as well as redundant basement membranes (Figs. 19.15 and 19.16a). Hypomyelination is the rule, with only 10–20 myelin lamellae per myelinated axon, regardless of its diameter (Dyck et al. 1971a). Ouvrier et al. (1987) reported that this represents the most consistent means of distinguishing morphologically between autosomal dominant CMT-1 and CMT-3: The g-ratio (see Chap.​ 3) was never above 0.69 in 10 cases of childhood-onset CMT-1 and never below 0.81 for 6 cases of Dejerine–Sottas. Myelin debris may be seen within Schwann cells. A slight increase in collagen pockets and denervated Schwann cell bands, subtle indicators of mild unmyelinated axon loss, has been reported (Ouvrier et al. 1987).

Since PMP duplications and point mutations are also associated with CMT-1A (vide supra), it is evident that the clinical and histological phenotype is dependent on additional factors, perhaps the site of mutation in the gene. Mutations in myelin P₀ protein have been found in patients satisfying clinical or histological criteria for DSS (Hayasaka et al. 1993d; Himoro et al. 1993). One of the three mutations occurred in the transmembrane domain of the P₀ protein and the others in the extracellular variable immunoglobulin-like region. In one instance these workers showed that P₀ was expressed in normal quantities in peripheral myelin (Tachi et al. 1994). As with the PMP-22 mutations, mutations in P₀ are also associated with the CMT-1 phenotype (vide supra: CMT-1B).

Of historical interest was the proposed grouping, by Gabreels-Festen and Gabreels (1993), of patients under the description of CMT type III with basal lamina onion bulbs (CMT-3-BLOB). Their review of the literature identified about 30 such cases (Harati and Butler 1985; Joosten et al. 1974; Lutschg et al. 1985; Balestrini et al. 1991; Ono et al. 1982; Moss et al. 1979; Vital et al. 1987; Vallat et al. 1987; Lyon 1969; Boylan et al. 1992; Guzzetta et al. 1982; Kennedy et al. 1977). Onset is most often at birth or infancy, with a variable disease severity ranging from arthrogryposis to survival to adult life with moderate disability. Conduction velocity is often below 6 m/s. Thus, these patients are clinically indistinguishable from the classic DSS syndrome discussed above.

Gabreels-Festen and colleagues suggested the delineation of a group of patients in which BLOBs are prominent, but which are distinguished histologically from the two groups previously mentioned by virtue of milder myelin abnormalities (Gabreels-Festen et al. 1992a; Gabreels-Festen and Gabreels 1993). These authors identified several similar cases in the literature (Meier et al. 1976; Nordborg et al. 1984; Smith et al. 1980). Clinically the patients suffer from a disease intermediate in severity between typical Dejerine–Sottas syndrome and CMT-1. Conduction velocities are usually in the 10–30 m/s range. All reported cases have been sporadic or autosomal recessive. Testing for the CMT-1A mutation has been negative (Gabreels-Festen and Gabreels 1993), but there may be overlap with CMT-4A and CMT-4C groups which have GDAP1 (ganglioside-induced differentiation-associated protein 1) and SH3TC2 gene defects identified.

Biopsy may show an increased endoneurial area with interstitial edema. There is reduction in myelinated fiber density, but unmyelinated fibers are relatively spared. Classic onion bulb formations are small and less frequent than BLOBs. Occasional myelin tomaculae are present. Teased fibers show segmental myelin changes without signs of active axonal degeneration. The major distinguishing point from the Dejerine–Sottas picture is that far fewer demyelinated axons are seen, and there is no evidence of the severe hypomyelination that characterized DSS. When measured, the g-ratio has been in the range 0.64–0.77.

The patients reported by Nordborg et al. (1984) showed intra-axonal lamellated inclusions in unmyelinated fibers and less often in myelinated fibers, as well as frequent redundant myelin loops.

The existence of yet another morphologically unusual variant of CDN, designated “autosomal recessive CMT with focally folded myelin (FFM)” was postulated by Gabreels-Festen and colleagues (Gabreels-Festen et al. 1990), who identified similar cases from the literature (Barbieri et al. 1994; Lutschg et al. 1985; Routon et al. 1991; Vital et al. 1987; Ouvrier et al. 1990 case 14.3; Nordborg et al. 1984). Umehara et al. 1993 described a dominantly inherited case of motor and sensory neuropathy with excessive myelin folding complex. Symptoms are present at birth or appear within the first year, with slow progression. Some are wheelchair bound in childhood, while others retain the ability to walk well into adulthood. Conduction velocities have been above and below the cutoff of 10–12 m/s usually used to delineated Dejerine–Sottas syndrome. CSF protein is usually normal.