Hepatitis C virus (HCV) subtypes were determined in 125 Iranian patients by phylogenetic analysis within the NS5B or 5'-UTR/core regions. Subtypes 1a and 3a were predominant accounting for 47 and 36%, whereas 1b and 4 accounted for 8 and 7%. This subtype distribution differs from that of Turkey and Pakistan, where subtypes 1b and 3a dominate and also from neighbouring Arabic countries where subtype 4 is the prevalent genotype. The Iranian 1a and 3a strains formed subclusters in the dendrogram indicating that these subtypes are indigenous to Iran. In contrast, the 1b strains intermixed with strains derived worldwide. Subtype 1a was frequent in South Iran (70%), while 3a was more prevalent in North-West Iran (83%), a region with a high proportion of Turkish inhabitants. Patients infected by blood products had more frequently subtype 1a (57%), while younger drug users had more frequently subtype 3a (54%). Genotype 4 was over-represented among haemodialysis patients in Tehran. One strain, most similar to genotype 5, was highly divergent in the NS5B region and further analysis is needed to assess the systematic status of this strain. In half of the patients with unknown source of infection only the 5'-UTR could be amplified, most of which were from North-West Iran and from patients younger than those with unknown source of infection with typable strains, mean age 29 versus 43 years. In conclusion, the NS5B sequence data revealed population based subtype patterns in Iran, the further study of which may help to understand the molecular epidemiology of HCV in a low-endemic area.