We read with interest the article by Fischer et al.1 evaluating the combined genotyping of interleukin (IL)28 single nucleotide polymorphisms (SNPs) rs12979860 and rs8099917, on pegylated-interferon (PegIFN) and ribavirin (RBV) treatment outcome in two large cohorts of 942 and 377 hepatitis C virus (HCV)-1-naïve patients. The authors found that in patients with a heterozygous variant for the nonresponder rs12979860 T allele, the additional genotyping of rs8099917 significantly improved sustained virological response (SVR) prediction, as carriers of the rs12979860CT/rs8099917TT genotype had a 55% SVR rate compared to the 40% seen in patients with the rs12979860CT/rs8099917TG genotype (P = 0.001). Based on these findings, the authors suggest the combined determination of rs12979860 and rs8099917 genotype as a further diagnostic procedure for improving treatment decisions.
To externally validate these findings, we analyzed the IL28B rs12979860 and rs8099917 in an Italian cohort of 187 naïve HCV-1 patients who received PegIFNalfa2a/alfa2b and RBV at standard doses and for standard durations, outside of any clinical trial. We performed single SNP and genotype analysis of data and found that: (1) The overall genotype distribution of IL28B rs12979860 CC, CT, and TT was 30%, 57%, and 12% and the distribution of rs8099917 TT, TG, and GG was 53%, 42%, and 5%, respectively; (2) SVR rates were 72%, 40%, and 30% for rs12979860 CC, CT, and TT and 60%, 37%, and 30% for rs8099917 TT, TG, and GG, respectively; (3) The most prevalent combined genotypes were: rs12979860CC/rs8099917TT (30%), rs12979860CT/rs8099917TT (22%), and rs12979860CT/rs8099917TG (35%). These prevalences are similar to those reported by Fisher et al. In our patients with a heterozygous genotype for rs12979860 the additional genotyping of rs8099917 did not improve SVR prediction. Namely, in our Italian cohort, carriers of the rs12979860CT/rs8099917TT genotype showed SVR rates similar to those seen in carriers of the rs12979860CT/rs8099917TG genotype (42% versus 39%). We can attribute these discrepant findings to differences in viral and host features between our patients and the ones analyzed in the Fisher et al. study. Indeed, Italian patients have often been infected in the 1970-1980s through parenteral exposure, and as a consequence our cohort was older (mean age 55 years), more often cirrhotic (24%), and more often infected by HCV-1b (90%) than HCV patients from other geographical regions such as northern Europe.2 Still, we cannot exclude that our relatively limited sample size might have prevented us seeing the differences observed by Fisher at al.; however, this only strengthens the concept that the combined determination of the rs12979860 and rs8099917 genotype may hold a strong predictive power for an SVR mainly in large cohorts of patients, such as those enrolled in drug development studies, but might be less relevant at the individual level in clinical practice.3