Genetic burden mirrors epidemiology of multiple sclerosis.

The prevalence of multiple sclerosis (MS) varies worldwide and, within Europe, decreases with latitude from Scandinavia to the Mediterranean populations. Several explanations have been offered for the prevalence gradient, including environmental factors such as sunlight exposure. However, differences in the frequency of genetic risk factors may well explain an important fraction of variation in prevalence of MS. 1 The Italian island of Sardinia, one of the highest risk regions worldwide with an MS prevalence of 224/100,000 inhabitants, is a notable exception on the north–south gradient and is characterized by a genetic background distinct from other European populations and long isolation. 2 In two papers in the current issue of Multiple Sclerosis Journal, Hadjixenofontos et al. and Barizzone et al. investigate the combined role of known MS genetic risk factors in this special population. 3,4 The current list of known MS risk variants includes several independent HLA alleles and 110 non-HLA variants, most of which have been identified in the context of the International Multiple Sclerosis Genetics Consortium, with contributions from continental Europe, the UK, USA and Australia. 5,6 Combining data from the current studies in Sardinia 3,4 with previous work in Europeans 5,6 and African Americans 7 demonstrates that common MS risk variants tend to be shared across populations even though their population frequencies can vary. Indeed, the vast majority (89%) of non-HLA risk variants detected in a mixed European population show over-representation of the risk allele also in Sardinian cases versus controls, with 42% reaching nominal significance as expected in a single study population of this size. 4 A summary measure, the multiple sclerosis genetic burden (MSGB), counts the total number of MS risk alleles a person carries, weighted by the risk effect size of each variant. As expected and as seen in other European populations, 8 the average genetic burden in patients is higher than in controls. 3,4 More remarkably, the genetic burden in Sardinian controls is significantly elevated compared with continental Italian or US controls and approximately equals the burden observed in patients in the latter populations. 3,4 This increased burden is not due to a few variants with exceptionally deviating frequencies in Sardinia but rather to a subtle increase in frequency for many risk variants. This is seen in the median difference of 1% between the frequency of non-HLA risk variants in Sardinia and continental Italy 4 or the average difference of two out of a …