In previous cross-sectional studies, we demonstrated that, in most patients with chronic hepatitis C, the composition and complexity of the circulating hepatitis C virus (HCV) population do not coincide with those of the virus replicating in the liver. In the subgroup of patients with similar complexities in both compartments, the ratio of quasispecies complexity in the liver to that in serum (liver/serum complexity ratio) of paired samples correlated with disease stage. In the present study we investigated the dynamic behavior of viral population parameters in consecutive paired liver and serum samples, obtained 3 to 6 years apart, from four chronic hepatitis C patients with persistently normal transaminases and stable liver histology. We sequenced 359 clones of a genomic fragment encompassing the E2(p7)-NS2 junction, in two consecutive liver-serum sample pairs from the four patients and in four intermediate serum samples from one of the patients. The results show that the liver/serum complexity ratio is not stable but rather fluctuates widely over time. Hence, the liver/serum complexity ratio does not identify a particular group of patients but a particular state of the infecting quasispecies. Phylogenetic analysis and signature mutation patterns showed that virtually all circulating sequences originated from sequences present in the liver specimens. The overall behavior of the circulating viral quasispecies appears to originate from changes in the relative replication kinetics of the large mutant spectrum present in the infected liver.Hepatitis C virus (HCV) is one of the leading causes of chronic liver disease worldwide (22). The quasispecies nature of the single-stranded RNA genome of HCV is thought to play a central role in maintaining and modulating viral replication (10, 29). In general, the natural history of HCV infection does not follow a defined pattern, but the rates of progression of the disease (from minimal changes of the liver to cirrhosis and hepatocarcinoma) vary greatly in different individuals (22). The cytopathic potential of the virus and the characteristics of the host's immune response have both been postulated to explain the observed differences in disease progression (6,17,18,45).In the absence of an appropriate animal model or culture system and in the context of the extreme heterogeneity of HCV, it has not been possible to associate particular sequences with distinct cytopathic potential. Similarly, attempts to correlate the process of quasispecies diversification with disease progression have yielded controversial results. While in crosssectional studies some authors found a correlation between quasispecies complexity and extent of liver damage (20,20,21,25,48), others did not (19,27,33,41). In longitudinal studies, the nucleotide complexity of the hypervariable domain of the HCV E2 region (HVR1) did not increase cumulatively over time but fluctuated in consecutive serum samples (2). Similarly, at more conserved regions of the genome, an oscillatory pattern of quasispecies complexit...