Diagnostic criteria for optic neuritis in the acute and subacute phase: clinical uses and limitations.

Abstract

A recent international consensus panel proposed diagnostic criteria for optic neuritis and a new classification. We aimed to investigate the clinical relevance of these diagnostic criteria and classification, in a cohort of patients hospitalized for a suspected diagnosis of optic neuritis. We included all patients hospitalized between 2017 and 2022 in our tertiary center for (sub)acute loss of visual acuity suggestive of optic neuritis. Clinical and paraclinical criteria obtained within the first 3months of symptoms were collected, as well as the final diagnosis which could be optic neuritis or non-optic neuritis. We constructed a contingency table comparing diagnoses based on physician experience to those based on the recently proposed criteria. The subtypes of optic neuritis based on the new classification were compared to subtypes based on the clinician experience. Two hundred fifty-seven patients were included in this study. Prevalence of optic neuritis in our cohort was 88.3%. Sensitivity and specificity of a correct diagnosis using the new criteria were, respectively, 99.5% and 86.7%. The proposed diagnostic criteria overdiagnosed four patients with optic neuritis and missed the diagnosis in one patient. According to the recent classification, idiopathic optic neuritis and clinical isolated syndrome were reclassified mainly as single isolated optic neuritis. In our specific cohort of patients hospitalized for acute and subacute optic neuropathy highly suspect of optic neuritis, we found that recently proposed diagnostic criteria and classification of optic neuritis are relevant for our clinical practice. Our interpretation of clinical requirement for definite and possible optic neuritis diagnosis might explain our excellent sensitivity and our high percentage of definite optic neuritis, relative to previous publications. The moderate specificity (86.7%) underlines the importance to include all contextual data in consideration for the diagnosis. The simplification of subgroups is useful, but our study highlights the complexity to find the adequate subgroup for seronegative NMOSD.