0631 Diagnostic Accuracy And Validity Of The Swiss Narcolepsy Scale For The Diagnosis Of Type 1 And Type 2 Narcolepsy Against Other Central Disorders Of Hypersomnolence

Abstract Introduction The Swiss Narcolepsy Scale (SNS) was previously reported in two independent studies to have a sensitivity-specificity of about 80–90% for the diagnosis of narcolepsy with cataplexy. The aim of this study was 1) to assess the validity and optimal cut-off point of the SNS in a third independent patients series in detecting type 1 (NT-1) and type 2 (NT-2) narcolepsy against other central disorders of hypersomnolence (CDH), 2) in order to develop a simplified form of the SNS. Methods Data from the Bern Sleep-Wake Registry were used. The validation of the diagnoses was based on the ICSD-3. A two-item simplified form (sf-SNS) was created from the SNS questions based on the discriminative capability (AUC) of the models including all possible combinations. For the estimation of the diagnostic accuracy of SNS and the sf-SNS we used sensitivity, specificity, Hosmer-Lemeshow goodness of fit test and the Brier score. In addition, we used logistic regression analysis to evaluate the scoring coefficients of the SNS. Results There were 299 individuals with CDH who completed the SNS scale, including 69 with NT-1, 16 with NT-2 and 214 with other CDH. Compared to the validator’s diagnosis, the accuracy in identifying NT-1 based on the SNS showed a sensitivity of 86% and a specificity of 88%, the Hosmer-Lemeshow goodness of fit test indicates sufficient calibration (p=0.700) but the Brier score was relatively high (0.87), indicating relatively high disagreement. Based on these data we recalculated the model coefficients for scoring SNS and improved sensitivity (93%), but reduced specificity to 82%. For the identification of NT-2 the specificity was 70% and the sensitivity 63%. The diagnostic accuracy for the sf-SNS in identifying NT-1 (sensitivity of 80% and specificity of 92%) was comparable to the SNS, while for NT-2, sensitivity was 44% and specificity was 83%. Conclusion The updated SNS and its simplified form sf-SNS are useful and valid screening tools for NT-1. Both scales show low sensitivity in detecting NT-2. Support (If Any):