We examined hepatitis C virus genotypes in 98 American patients with chronic hepatitis C virus infection by means of two methods; restriction fragment length polymorphism analysis and line probe assay, which is based on type-specific sequence variations in the 5' untranslated region. Type 1 was present in 73 patients (74%), type 2 in 15 (15%), type 3 in 6 (6%) and type 4 in 1 (1%). Line probe assay further subdivided type 1 into 1a (n = 35) and type 1b (n = 37) and type 2 into type 2a (n = 6) and 2b (n = 9). Two patients (2%) had both restriction fragment length polymorphism and line probe assay evidence of dual infection (with types 1 and type 2) while another case had both type 1a and 1b by line probe assay. One patient was untypable by either technique. There was no correlation between infecting genotype and presumed cause, serum indexes of necroinflammatory activity, or age or sex of the patients studied or known duration of infection. Patients with type 2 hepatitis C virus had more severe liver disease histologically (p = 0.0027) compared with other genotypes but, paradoxically, had significantly lower levels of circulating hepatitis C virus RNA (12.1 +/- 12.8 x 10(5) genome equivalents/ml) than other types (36.4 +/- 44.8 x 10(5) genome equivalents/ml, p < 0.001). Response to interferon was less likely to be sustained in patients infected with type 1 than in those infected with other types (7% vs. 40%, p < 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)