Application des critères de de Seze et al. sur une série de 14 patients ayant présenté un premier événement démyélinisant aigu sévère

Introduction. - The application of de Seze et al. criteria (2007) to patients presenting a first severe acute demyelinating event helps to distinguish acute disseminated encephalomyelitis (ADEM) from other CNS inflammatory diseases, with 83% sensitivity and 95% specificity. We applied these criteria to 14 patients who presented a first severe acute demyelinating event and whose later clinical course enabled clear identification of the neurological diagnosis. Method/Patients. - This study concerned 14 patients who presented a first acute demyelinating event. Initially, there were 16 patients but two were excluded because their initial clinical condition (isolated acute retrobulbar optic neuritis in one and acute cervical myelitis in the other) would have excluded them in the princeps article. We identified 11 women (78.6%) and three men (21.4%) with a mean age of 33.7 +/- 12.5 years. Follow-up ranged from three months to 11.5 years after the initial episode (average four years). At last follow-up, the diagnosis was ADEM in seven patients (50%) and multiple sclerosis (MS) in seven (50%). Five of seven patients in the MS group had a tumor-like presentation (71.4%), this parameter partly explaining the initial discrepancy in diagnosis. When applied to our series, de Seze criteria for ADEM exhibited 85.7% sensitivity and 71.4% specificity. Conclusion. - Applying the new criteria, we did not find the same sensitivity, specificity, and positive and negative predictive values as in the original article. The lack of specificity arose from the misclassification of MS patients with a tumor-like presentation (two out of five false negatives). One of the explanations is that the clinical criteria used can be part of atypical forms of MS, in particular in its tumor-like presentation. De Seze et al. criteria can be an invaluable help for the clinician in the diagnosis of a first severe demyelinating event. Considering our results, these criteria should not be applied for patients with a tumor-like form of MS. A prospective study in a larger cohort is needed to confirm or invalidate these preliminary results. (C) 2008 Publie par Elsevier Masson SAS.