It is generally accepted that while inflammatory demyelinating neuropathies often cause patchy demyelination resulting in conduction block, temporal dispersion and variation in conduction velocities, demyelinating hereditary neuropathies such as Charcot–Marie–Tooth (CMT) disease type 1A are usually characterised by homogeneous slow conduction.

X-linked CMT is caused by mutations in the gap junction beta 1 (GJB1) gene encoding connexin 32, a gap junction protein, resulting in intermediate conduction velocities. At our institution, mean median nerve conduction velocity in men with GJB1 mutations is 33.2 ± 7.5 m/s (n = 13) and 47.7 ± 7.0 m/s in women (n = 8), similar to other recent series.1 2 There is increasing evidence that nerve conduction is not always homogeneous in GJB1 associated CMT.1^–4 Here we present three cases in which the neurophysiology suggested an inflammatory demyelinating neuropathy, but who failed to respond to treatment and were subsequently found to have mutations in GJB1.

Case reports

Case No 1

A 23-year-old man presented with parasthesia affecting both feet that progressed to mild hand parasthesia and weakness over 6 months. Although slightly clumsy, he had been good at sport as a teenager. He had minimal wasting but normal power in his hands, he was areflexic and had reduced vibration and pinprick distally. CSF was acellular, with a raised protein of 0.78 g/l. His initial nerve conduction studies (NCS) (done elsewhere) suggested chronic inflammatory demyelinating polyneuropathy (CIDP) (unavailable). Anti-GM1 antibodies were negative. He was given intravenous immunoglobulin, with a mild subjective effect on the sensory symptoms, but no objective effect. Eight years after presentation he was referred to our institution for a second opinion, where no significant progression of his disease was documented. Repeat NCS showed a patchy demyelinating neuropathy with mild temporal dispersion with proximal stimulation and probable conduction block in the left median nerve (table 1). Given the disease duration and stability, the chromosome 17 duplication was screened (negative) followed by the GJB1 gene, revealing a novel Val91Leu (GTG/CTG) mutation. All family members were reportedly asymptomatic but none was available to be seen.

Case No 2

A 44-year-old woman gave a 15 year history of slowly progressive wasting and weakness of the left hand associated with mild numbness of the ulnar border of the left hand. Examination at the age of 34 years revealed moderate weakness in both the left ulnar and left median nerve distributions. NCS suggested multifocal motor neuropathy, with sensory responses almost completely normal but conduction block in the left median nerve between the wrist and elbow (table 1). NCS in all other limbs were normal. Trials of intravenous immunoglobulin and azathioprine were unsuccessful.

Repeat NCS at age 44 years was similar (conduction block in the left median nerve) and CSF protein was 0.39 g/l. The chromosome 17 duplication (negative), myelin protein zero (negative) and GJB1 genes were sequenced as part of a study of genetic influences in inflammatory neuropathies, revealing a novel GJB1 Asn226Ser (AAT/AGT heterozygote) mutation. All family members were reportedly asymptomatic but none was available to be seen.

Case No 3

A 45-year-old woman first presented to another hospital 15 years previously with mild weakness of her hands and feet. She recalled sprained ankles as a child and occasional falls in her twenties. NCS at that time (now unavailable) apparently supported a diagnosis of CMT1 but chromosome 17 duplication was negative. Over 2 years in her mid-thirties, she noted a clear deterioration, with increased weakness and peripheral parasthesia. Examination revealed distal weakness, hyporeflexia and reduced distal sensation. Repeat NCS showed mild temporal dispersion and probable conduction block in the ulnar nerve in the forearm (table 1), suggesting superimposed CIDP. CSF protein was normal with unmatched oligoclonal bands. MRI brain scan was normal. A sural nerve biopsy showed an axonal neuropathy with some thinly myelinated nerve fibres and no significant inflammation. A superimposed CIDP on a background of CMT1 was diagnosed. Despite numerous attempts at treatment, including oral steroids, intravenous immunoglobulin, azathioprine and plasmapheresis, her condition remained stable but did not improve over several years, with persistence of conduction block. Given the disease duration and stability, the GJB1 gene was sequenced, revealing a Val91Met (GTG/ATG) mutation. All family members were reportedly asymptomatic but none was available to be seen.

Discussion

These three cases highlight the possibility that patients with GJB1 mutations might be erroneously diagnosed as having an inflammatory neuropathy. The patients all presented with an isolated neuropathy (no CNS involvement) and in each one the NCS suggested an inflammatory neuropathy. While a superimposed inflammatory neuropathy cannot be excluded in any of the patients, the lack of response to treatment and minimal progression over time would be surprising.

Other groups have reported isolated cases with GJB1 mutations in whom NCS have shown heterogeneous motor conduction velocities in a single nerve as well as between different nerves in an individual.3 4 Furthermore, there are reports of conduction block,1 temporal dispersion2^–4 or abnormal terminal latency index,4 which are all more characteristic of the patchy pathology seen in inflammatory conditions than the more homogeneous findings expected in hereditary neuropathies.

If, as seems likely, the GJB1 mutations are pathogenic in these cases, the cause of these findings is unknown. Segmental demyelination has been noted in pathological specimens of a small proportion of patients with X-linked CMT. It is well known that probands with the same GJB1 mutation can show differing conduction velocities.2 Although some variability in females may relate to different lyonisation of the X chromosome in Schwann cell precursors,1 this cannot explain observations in the male population. It is interesting to note that conduction block has rarely been reported in other intermediate CMT neuropathies (eg, a man with a S140T mutation in the Myelin protein zero gene).5 Furthermore, temporal dispersion has been reported in patients with CMT1C with missense mutations in the small integral membrane protein of lysosome/late endosome (SIMPLE) gene.6

Given our findings it seems prudent to consider sequencing the GJB1 gene in suspected cases of inflammatory demyelinating neuropathy which fail to respond to treatment even if there is temporal dispersion and conduction block present.

We acknowledge the Medical Research Council and the Muscular Dystrophy Campaign for funding support. This work was undertaken at University College London Hospitals/University College London, which received a proportion of funding from the Department of Health’s National Institute for Health Research Biomedical Research Centres funding scheme.

Acknowledgements

We thank Dr PB Kang, Boston, for discussion.