Arising from: Hemmer B and Stüve O (2007) Revised criteria for neuromyelitis optica–a new diagnostic standard? Nat Clin Pract Neurol 3: 132–133 doi:10.1038/ncpneuro0420

We have read with interest the commentary by Bernhard Hemmer and Oliver Stüve regarding the recent proposal by Wingerchuk et al.1 to include NMO-IgG, a newly detected serum reactivity, in the diagnostic criteria for neuromyelitis optica (NMO). We share the commentary authors' concerns regarding the fact that this proposal was made prior to independent confirmation of the diagnostic sensitivity and specificity of the test.

Your readers might, however, like to know that the sensitivity and specificity of NMO-IgG has now been assessed by our laboratory in Oxford by means of indirect immunofluorescence, using the method of Lennon et al. as described in their patent application. We detected NMO-IgG in 22/36 NMO patients and 4/5 patients with isolated longitudinally extensive transverse myelitis (LETM), but only in 1/137 control samples, which included samples from 80 patients with multiple sclerosis (MS). Our results (63% sensitivity and 99% specificity) are well in line with previous findings by the Mayo group (58–73% sensitivity and >90% specificity).1,2,3 In another study in Berlin, we demonstrated and quantified antibodies to aquaporin-4 (Aqp-4), the water channel proposed as the antigenic target of NMO-IgG by Lennon et al.,4 using a newly developed radioimmunoprecipitation assay. We found Aqp-4 antibodies in 27/43 (63%) patients with NMO/LETM, but only in 5/291 (1.7%) controls including patients with other inflammatory and neurological disorders.5 Our findings independently confirm that NMO-IgG antibodies, and more specifically Aqp4 antibodies, are a marker for NMO, suitable for differentiation between NMO/LETM and other autoimmune diseases of the CNS including MS. Thus, NMO-IgG/Aqp-4 antibodies may well be taken as a 'supportive criterion' for the diagnosis of NMO.

The diagnosis can still be based on clinical and MRI criteria as demanded by the authors in their final statement. Yet, the inclusion of Aqp-4 antibodies increases the diagnostic specificity in NMO/LETM patients. More importantly, in patients with non-LETM myelitis not fulfilling MRI criteria for MS as well as in those with longitudinal spinal cord lesions and MS-like brain lesions, NMO-IgG or Aqp-4 antibodies are the only parameter that so far allows discrimination between MS and NMO with high specificity. The vertical extension of spinal cord lesions in NMO is subject to timing issues, because lesions might develop over the first days after onset; in addition, large areas of high signal intensity mimicking NMO can occur in more advanced MS when focal lesions merge.1,3 In these patients, NMO-IgG or Aqp-4 antibodies, which are frequently positive already at first presentation (S Jarius, D Franciotta and A Vincent; unpublished data), represent a nonredundant and, therefore, important disease marker.