To study longitudinal survival probability by period of MS-diagnosis and time from disease onset to diagnosis using nation-wide data from the Nordic MS-registries.

All patients in the MS registries in Norway (N=11869), Sweden (N=20696), Denmark (N=21492) and Finland (N=17716) alive at the study starting point of Jan 1^st 2000 were included in the study, with a follow-up time for up to 20 years. A total of 6333 deaths were recorded during the follow-up period ending Dec. 31^st 2020. Long-term survival in MS patients was analyzed in each MS-registry cohort using Cox proportional hazards regression. Covariates included in the regression model were sex (ref:male), age at onset, calendar period of diagnosis (< 1980, 1981-1989, 1990-1999 (ref), >1999), MS phenotype at diagnosis (relapsing-remitting MS (ref), primary progressive MS) and diagnostic delay (time from MS onset to MS diagnosis in years: <2 (ref), 2-4, 5-10, > 10). Results were then combined and used in meta-analyses to identify common effects or variations as well as overall estimates using all four MS registries.

The overall estimated values for hazard ratios and 95% confidence intervals for sex were 0.66 [0.61, 0.71], for onset age 1.06 [1.05, 1.07], for MS phenotype 1.65 [1.44, 1.89], for period of diagnosis 4.95 [2.44, 10.07], 2.03 [1.49, 2.76], 0.48 [0.35, 0.65], and for diagnostic delay 1.37 [1.23, 1.52], 1.70 [1.55, 1.86], 3.52 [3.09, 4.00]. The estimated effects were all significant, while the test for heterogeneity was significant for diagnosis period and diagnostic delay.