74 Chapter 3 Table 2: Hypersomnolence disorder and influenza incidence rates Estimate Confidence interval (95%) p-value NT1 and type A H1N1 All subjects 0.0017 [0.0011, 0.0022] <.0001 Children 0.0008 [0.0004, 0.0013] 0.0001 Adults 0.0008 [0.0006, 0.001] <.0001 NT1 and type A H3N2 All subjects -0.0034 [-0.006, -0.0008] 0.005 Children -0.002 [-0.004, -0.0003] 0.011 Adults -0.0014 [-0.0024, -0.0001] 0.009 NT2+IH and type A H1N1 All subjects 0.0005 [0.0002, 0.0008] 0.0007 Children 0.0004 [0.0002, 0.0005] <0.0001 Adults 0.0002 [0.0001, 0.0004] 0.0738 NT2+IH and type A H3N2 All subjects -0.0009 [-0.0019, 0.0004] 0.1308 Children -0.0007 [-0.0014, -0.00006] 0.0341 Adults -0.0002 [-0.0007, 0.0009] 0.9567 Results of the linear mixed model analyses analysing the relationship between hypersomnolence disorder incidence rates and preceding influenza season severity combining data from the four largest samples (France, Italy, the Netherlands and Czech Republic). IH = idiopathic hypersomnia; NT1 = narcolepsy type 1; NT2 = narcolepsy type 2. Per country correlational analyses revealed that the type A influenza effects were strongest in the Netherlands and France. In the Netherlands we additionally identified a strong positive correlation with type B Victoria influenza severity in children (correlation coefficient 0.94, P-value < 0.001, as shown in Figure 4). When shifting incidence rates of influenza strains one year backward as a sensitivity analysis, no significant correlations remained present between type A H1N1, type A H3N2 or type B Victoria, and NT1 incidence rates (Appendix B, Supplementary Figure 16).
RkJQdWJsaXNoZXIy MjY0ODMw