Seasonal variation of relapse rate in multiple sclerosis is latitude dependent

Objective Previous studies assessing seasonal variation of relapse onset in multiple sclerosis have had conflicting results. Small relapse numbers, differing diagnostic criteria, and single region studies limit the generalizability of prior results. The aim of this study was to determine whether there is a temporal variation in onset of relapses in both hemispheres and to determine whether seasonal peak relapse probability varies with latitude. Methods The international MSBase Registry was utilized to analyze seasonal relapse onset distribution by hemisphere and latitudinal location. All analyses were weighted for the patient number contributed by each center. A sine regression model was used to model relapse onset and ultraviolet radiation (UVR) seasonality. Linear regression was used to investigate associations of latitude and lag between UVR trough and subsequent relapse peak. Results A total of 32,762 relapses from 9,811 patients across 30 countries were analyzed. Relapse onset followed an annual cyclical sinusoidal pattern with peaks in early spring and troughs in autumn in both hemispheres. Every 10° of latitude away from the equator was associated with a mean decrease in UVR trough to subsequent relapse peak lag of 28.5 days (95% confidence interval = 3.29–53.71, p = 0.028). Interpretation We demonstrate for the first time that there is a latitude‐dependent relationship between seasonal UVR trough and relapse onset probability peak independent of location‐specific UVR levels, with more distal latitude associated with shorter gaps. We confirm prior meta‐analyses showing a strong seasonal relapse onset probability variation in the northern hemisphere, and extend this observation to the southern hemisphere. Ann Neurol 2014;76:880–890