Three families with homozygous or compound heterozygous MFN2 mutations were examined (the authors do not describe how many were screened to discover these families). The third patient described is homozygous for this variant (R707W).
The summary currently presented in OMIM (as of 1/2013) appears to be mistaken in its understanding of the paper. It states: “Each parent was heterozygous for the mutation and showed a mild form of peripheral neuropathy. Nicholson et al. (2008) noted that the apparent autosomal recessive inheritance of the disorder in this family was actually semidominant and that the parents showed incomplete penetrance.”
In fact, the authors discuss the combined effect of the two mutations in an attempt to explain why the parents were not symptomatic despite carrying mutations: they attempt an “explanation for the parents being asymptomatic or very mildly symptomatic despite carrying heterozygous mutations”. Elsewhere the parents are all described as “asymptomatic” and with “minimal clinical findings of neuropathy and minor electrophysiologic evidence of neuropathy”. This is all consistent with arguing for a recessive mode of inheritance.
A survey of of 107 unrelated patients with type 2 Charcot-Marie-Tooth disease and 43 “median” cases (type 1, but close to the threshold motor nerve conduction velocity used to separate types). Three patients from one family were found to be compound heterozygous for this variant and G108R. The authors report that the parents (each carrying one of these variants) show no symptoms of neuropathy and had normal electroneuromyographic findings.