257 Clustering in Central Disorders of Hypersomnolence Current diagnosis and centres of inclusion Physicians were, in general, most confident in diagnosing narcolepsy type 1 and more doubtful with narcolepsy type 2 and idiopathic hypersomnia (Figure 3A). Clusters 1 through 4 included mainly people diagnosed with narcolepsy type 1 with a definite diagnostic certainty (91%, 75%, 65%, and 83%). Contingency table statistics for narcolepsy type 1 within clusters 1 through 4 showed a sensitivity of the clusters of 0.76, specificity of 0.83, positive predictive value of 0.93, and negative predictive value of 0.76. Within cluster 3, there were 18 individuals (20%) without cataplexy but with sleep drunkenness, hypnagogic hallucinations, and often sleep paralysis. Clusters 5 and 6 were dominated by people without cataplexy, and narcolepsy type 2 and idiopathic hypersomnia diagnoses were similarly mixed over the 2 clusters (cluster 5, 45% narcolepsy type 2 and 40% idiopathic hypersomnia; cluster 6, 45% narcolepsy type 2 and 27% idiopathic hypersomnia). Contingency table statistics for narcolepsy type 2 and idiopathic hypersomnia within clusters 5 and 6 showed a sensitivity of the clusters of 0.76, specificity of 0.91, positive predictive value of 0.78, and negative predictive value of 0.90. There was also a considerable proportion of individuals with narcolepsy type 1 in clusters 5 and 6 (cluster 5, 15%; cluster 6, 28%). Most of these individuals had atypical or mild cataplexy (cluster 5, 35%; cluster 6, 38%) or absent cataplexy but low hypocretin-1 levels (cluster 5, 26%; cluster 6, 33%). Cluster 7 was a mixed cluster of mainly narcolepsy types 1 and 2. Centres of inclusion were evenly spread over the clusters (Figure 3B). In Appendix B, we present summary measures of the population using current ICSD-3 diagnoses. 9
RkJQdWJsaXNoZXIy MjY0ODMw