We have performed a replication study based on the recent findings of the first published GWAS on CH. The three most significant variants,
rs12668955,
rs1006417 and
rs147564881 were selected and genotyped in a Swedish CH case-control study population. We found no association for either genotypes or alleles with CH for the common SNPs, and further we found no CH patient carrying the mutated allele of
rs147564881. The discrepancy between our study and the discovery study can have several reasons. The control materials exhibit considerable differences. The Italian GWAS used 360 cigarette smokers, while we used anonymous blood donors, 14% of which are presumably smokers, as this is the proportion of smokers in the Swedish population. Moreover, there is a significant difference in gender ratio between the Swedish and the Italian material. The Italian cohort have an unusually high proportion of males (84%) compared to the Swedish cohort (68,3%). As a control experiment, we therefore verified our results in a gender stratified analysis conducted with male subjects only, successfully replicating a lack of association. Also, there might be differences in genetic background between the Swedish and Italian populations. The minor allele frequencies (MAF) for the two common SNPs differs between our study and the Italian report, in particular for
rs1006417 which is protective in the Italian study. In our material, although the difference is non-significant, cases have a higher MAF (0.245) than controls (MAF 0.216) which would be indicative of an increased rather than decreased risk. The association discovered in the Italian cohort might also reflect a linkage disequilibrium (LD) between these SNPs and other genetic variations truly linked to disease that are not present in the Swedish population. Since we did not genotype any additional marker at these loci, we could not control for a potential difference in the LD structure between the Swedish and Italian populations. Another limitation of our study is the possibility that other rare
MME mutations might be associated to CH in our material. Last, life style and environmental factors, e.g. seasonal variation has been reported to influence CH, smoking is a risk factor for CH, other plausible factors might be inflammations and diet, [
7,
8,
12,
13,
17], and gene-environment interactions might cause genetic risk factors to differ in specific populations or various parts of the world.